Data processing method and device based on large language model

Abstract

The embodiment of the invention provides a data processing method and device based on a large language model, and the method comprises the steps: obtaining a standard service copywriting of a service direction in a process that a user accesses a resource service and interacts with a service party, the standard service copywriting is written into a copywriting processing template and an object recognition template to obtain a copywriting processing text and an object recognition text, and further, the copywriting processing text and the object recognition text are input into a large language model to be processed to obtain structured data and resource objects; and finally, the structured data and the object data of the resource object are sent to the user terminal, and the user terminal performs rich text label conversion and rich text display on the structured data and the object data, so that the interactive content of the service party is visually displayed in a rich text form on the user terminal.

Description

Technical Field

[0001] This document relates to the field of data processing technology, and in particular to a data processing method and device based on a large language model. Background Art

[0002] With the continuous development and promotion of the Internet and artificial intelligence, many services can be automated through online large language models. For example, in the scenario where users access online services, online services can provide targeted services to users by calling large language models. However, as more and more online services are processed based on large language models, the competition faced by all parties is becoming increasingly fierce. In this case, higher requirements are placed on service processing through large language models. Summary of the invention

[0003] One or more embodiments of the present specification provide a data processing method based on a large language model, including: obtaining a standard service copy for a service provider of a resource service to interactively respond to a user. Writing the standard service copy into a copy processing template to obtain a copy processing text, and writing the standard operation copy into an object recognition template to obtain an object recognition text. Inputting the copy processing text into a large language model for copy processing and structural processing to obtain structured data, and inputting the object recognition text into a large language model for resource object recognition to obtain a resource object. Sending the structured data and the object data of the resource object to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

[0004] One or more embodiments of the present specification provide another data processing method based on a large language model, including: obtaining a standard service text, structured data, and object data of a resource object for a service provider of a resource service to interactively respond to a user. The structured data is obtained by calling a large language model for text processing and structural processing. The resource object is obtained by calling a large language model for resource object identification. Corresponding rich text tags are generated based on the structured data and the object data, and the corresponding text fields in the standard service text are updated based on the rich text tags to obtain rich text. The rich text is visually displayed in the interactive interface with the service provider.

[0005] One or more embodiments of the present specification provide a data processing device based on a large language model, including: a text acquisition module, configured to obtain a standard service text for the service direction of a resource service to interactively respond to a user. A text generation module, configured to write the standard service text into a text processing template to obtain a text processing text, and write the standard operation text into an object recognition template to obtain an object recognition text. A text processing module, configured to input the text processing text into a large language model for text processing and structural processing to obtain structured data, and input the object recognition text into a large language model for resource object recognition to obtain a resource object. A data sending module, configured to send the structured data and the object data of the resource object to a user terminal, so as to convert the structured data and the object data into rich text tags and display them in rich text.

[0006] One or more embodiments of the present specification provide another data processing device based on a large language model, including: a data receiving module, configured to obtain standard service text, structured data, and object data of resource objects for the service party of the resource service to interactively respond to the user. The structured data is obtained by calling the large language model for text processing and structural processing, and the resource object is obtained by calling the large language model for resource object identification. A label generation module is configured to generate corresponding rich text labels according to the structured data and the object data, and update the corresponding text fields in the standard service text based on the rich text labels to obtain rich text. A rich text display module is configured to perform a visual display of the rich text in the interactive interface with the service party.

[0007] One or more embodiments of the present specification provide a data processing device based on a large language model, including: a processor; and a memory configured to store computer-executable instructions, wherein when the computer-executable instructions are executed, the processor: Obtain the standard service copy for the service provider of the resource service to interactively respond to the user. Write the standard service copy into the copy processing template to obtain the copy processing text, and write the standard operation copy into the object recognition template to obtain the object recognition text. Input the copy processing text into the large language model for copy processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain the resource object. Send the structured data and the object data of the resource object to the user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

[0008] One or more embodiments of the present specification provide another data processing device based on a large language model, including: a processor; and a memory configured to store computer-executable instructions, wherein when the computer-executable instructions are executed, the processor: Obtain the standard service copy, structured data and object data of the resource object for the service provider of the resource service to interactively respond to the user. The structured data is obtained by calling a large language model for copy processing and structural processing. The resource object is obtained by calling a large language model for resource object identification. Generate corresponding rich text tags based on the structured data and the object data, and update the corresponding copy fields in the standard service copy based on the rich text tags to obtain rich text. Visually display the rich text in the interactive interface with the service provider.

[0009] One or more embodiments of the present specification provide a computer-readable storage medium for storing computer-executable instructions, which implement the following process when executed: Obtain a standard service copy for the service provider of a resource service to interactively respond to a user. Write the standard service copy into a copy processing template to obtain a copy processing text, and write the standard operation copy into an object recognition template to obtain an object recognition text. Input the copy processing text into a large language model for copy processing and structural processing to obtain structured data, and input the object recognition text into a large language model for resource object recognition to obtain a resource object. Send the structured data and the object data of the resource object to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

[0010] One or more embodiments of the present specification provide another computer-readable storage medium for storing computer-executable instructions, which implement the following process when executed: obtaining standard service text, structured data, and object data of resource objects for interactive responses from a service provider of a resource service to a user. The structured data is obtained by calling a large language model for text processing and structural processing. The resource object is obtained by calling a large language model for resource object identification. Corresponding rich text tags are generated based on the structured data and the object data, and the corresponding text fields in the standard service text are updated based on the rich text tags to obtain rich text. The rich text is visually displayed in the interactive interface with the service provider. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate one or more embodiments of the present specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present specification. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor. Figure 1 A schematic diagram of an implementation environment of a data processing method based on a large language model provided for one or more embodiments of this specification; Figure 2 A processing flow chart of a data processing method based on a large language model provided for one or more embodiments of this specification; Figure 3 A schematic diagram of an interactive interface provided for one or more embodiments of this specification; Figure 4 A processing flow chart of a data processing method based on a large language model applied to a first resource service scenario provided in one or more embodiments of this specification; Figure 5 A processing flow chart of a data processing method based on a large language model applied to a second resource service scenario provided in one or more embodiments of this specification; Figure 6 A processing flow chart of a data processing method based on a large language model provided for one or more embodiments of this specification; Figure 7 A schematic diagram of an embodiment of a data processing device based on a large language model provided for one or more embodiments of this specification; Figure 8 A schematic diagram of another data processing device embodiment based on a large language model provided for one or more embodiments of this specification; Fig. 9 A schematic diagram of the structure of a data processing device based on a large language model provided for one or more embodiments of this specification; Fig.10 A schematic diagram of the structure of another data processing device based on a large language model provided for one or more embodiments of this specification. DETAILED DESCRIPTION

[0012] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0013] The data processing method based on the large language model provided in one or more embodiments of this specification can be applied to the implementation environment of resource services. Figure 1 , the implementation environment at least includes: User terminal 101, server 102 and large language model 103; The user terminal 101 is used to access the resource service. During the access process, the user can interact with the service provider of the resource service. The user terminal 101 is also used to receive and display data sent by the server 102 during the interaction process. The user terminal 101 can specifically be a mobile phone, a personal computer, a tablet computer, an e-book reader, a device for information interaction based on VR (Virtual Reality, virtual reality technology), a vehicle terminal, an IoT device, a wearable smart device, a laptop portable computer, a desktop computer, etc.; Server 102 is used to process interactive responses when users access resource services and interact with service providers, specifically including processing quasi-service documents for interacting with users on the service side, and sending the processed document data to user terminal 101; server 102 can be a single server, or a server cluster composed of several servers, or one or more cloud servers in a cloud computing platform.

[0014] The large language model 103 is used to perform text processing according to the input text, and to perform object recognition according to the input text. The large language model 103 can run on the server 102 , and can also run on a service platform outside the server 102 .

[0015] In this implementation environment, when a user interacts with a service provider when accessing a resource service, the server 102 obtains a standard service copy that the service provider uses to interact with the user, writes the standard service copy into a copy processing template and an object recognition template respectively to obtain a copy processing text and an object recognition text, and then inputs the copy processing text into the large language model 103 for copy processing and structural processing to obtain structured data, and inputs the object recognition text into the large language model 103 for resource object recognition to obtain a resource object, and after obtaining the object data of the resource object, sends the structured data and the object data of the resource object to the user terminal 101, and the user terminal 101 generates a corresponding rich text tag according to the received structured data and the object data of the resource object, and updates the standard service copy based on the rich text tag to obtain rich text, and finally displays the rich text in the interactive interface with the service provider, thereby enhancing the text of the standard service copy for interaction between the service provider and the user with the help of the large language model 103, and being able to visually display the enhanced standard service copy in rich text form on the user terminal 101.

[0016] It should be noted that, considering that the user's resource access records and other related data involved in this specification may belong to the user's privacy to a certain extent, if you want to collect the user's resource access records and other related data, you can obtain the user's authorization before collecting the data, so that the operation of collecting data complies with relevant data management regulations. For example, data authorization can be performed when the user accesses the resource service for the first time, or during any time the user accesses the resource service, or when the user triggers the service identification. The specific method of data authorization can be to send a user data authorization reminder to the user, and the user can obtain the user's data authorization after confirming the reminder through an instruction, or the method of data authorization can also be to obtain the user's data authorization by signing a data authorization agreement.

[0017] One or more embodiments of a data processing method based on a large language model provided in this specification are as follows: Reference Figure 2 The data processing method based on the large language model provided in this embodiment specifically includes steps S202 to S208.

[0018] Step S202: obtaining a standard service document for interactively responding to a user by a service provider of a resource service.

[0019] The resource service described in this embodiment refers to the resource-related services provided within the application, such as the resource services provided by the application for recommending resource items, trading, and other related processing, where the resource items can be resource items for resource management, and the managed resources can be equity resources, such as asset types that can be used for trading (funds, stocks, bonds), etc. The managed resources can also be financial resources or virtual resources, and the virtual resources can be virtual resources such as carbon savings and points.

[0020] The service provider of a resource service refers to the entity assigned to the user to provide services during the access process of the resource service. The service provider may be the service personnel of the service provider, such as customer service personnel or professional service personnel. The service provider may also be a virtual customer service. The service provider may be assigned to the user by the service platform of the resource service. The service platform may be the operator, provider or developer of the application. It may also be assigned to the user by the resource provider or resource transaction party of the resource service. Specifically, the service provider may be assigned to the user if the user data and / or resource data of the user meets preset conditions. For example, a service personnel may be assigned to the user if the user's resource access data and / or resource transaction data of the resource service meets preset conditions. For another example, a service personnel may be assigned to the user if the user's level in the application is higher than the preset level.

[0021] In this embodiment, in order to improve the service experience of users in the process of accessing resource services, in the process of interaction between the service provider of resource services and users, a standardized service process (standard service process) can be set for the service provider, and a standardized service document, i.e., a standard service document, can be set for each service node in the standard service process; in the specific implementation process, in the scenario where the service provider of resource services interacts with users, in the process of determining the standard service document for the service provider to respond to the user's interaction, a certain service node (target node) standard service document can be selected in the service document according to the service document of the standard service process as the corresponding standard service document for interacting with the user; In addition, the standard service copy can be determined by matching the user's resource access record with the standard service process of the resource service, that is, determining the standard service copy that matches the resource access record in the standard service process of the resource service; or, the standard service copy can be determined based on the target node selected by the service provider in the standard service process of the resource service, that is, using the standard service copy of the target node selected by the service provider in the standard service process of the resource service as the standard service copy for the service provider to interact with the user.

[0022] Optionally, the service party interactively responds to the user after the service party triggers the target node in the standard service process of the resource service, or after matching the user's resource access record with the standard service process of the resource service to obtain the target node.

[0023] Correspondingly, the standard service copy can be determined after matching the user's resource access record with the standard service process of the resource service, or it can also be determined after the service provider triggers the target node in the standard service process of the resource service, that is: determine the standard service copy based on the target node triggered by the service provider.

[0024] For example, when a user interacts with the service provider during the process of accessing a resource service, the service provider selects the target node in the standard service process of the resource service. The standard service copy based on the target node is "If you can accept a six-month closed period, it is recommended that you consider the xxxx equity resource project. The project was established on xx / xx / xxxx. Since its establishment, the annualized rate of return has been 3.07%, the rate of return in the past year has been 2.45%, the maximum drawdown is 0.45%, and the maximum drawdown repair time is 10 days. The quarterly reports since its establishment show that the project's rate of return has been positive for the past 6 months."

[0025] Step S204, writing the standard service copy into a copy processing template to obtain a copy processing text, and writing the standard operation copy into an object recognition template to obtain an object recognition text.

[0026] The copy processing text described in this embodiment refers to the input text / model input for input into the large language model for processing, that is, prompt. Specifically, it is the copy processing text that performs corresponding processing on the standard service copy by calling the large language model. The copy processing template refers to the template used to generate the copy processing text, that is, prompt template. The copy processing template contains the task text for copy processing. The copy processing template contains the task text for copy processing. The task text for copy processing includes the task text for extracting key content and also includes the task text for extracting resource recommendations. Optionally, the copy processing template includes: the task text for extracting key content and / or, the task text for extracting resource recommendations.

[0027] The object recognition text refers to the input text / model input used to input the large language model to perform object recognition on the standard service text, that is, prompt. The object recognition template is a template used to generate object recognition text, that is, prompt template.

[0028] In the specific implementation, the standard service copy is written into the copy processing template and the object identification template respectively to obtain the copy processing text and the object identification text. Specifically, the standard service copy is written into the copy processing template to obtain the copy processing text, and the standard operation copy is written into the object identification template to obtain the object identification text. It should be pointed out that the processing sequence of writing the standard service copy into the copy processing template to obtain the copy processing text and the processing sequence of writing the standard operation copy into the object identification template to obtain the object identification text are not limited in the execution process. The processing sequence of writing the standard service copy into the copy processing template to obtain the copy processing text can be executed first, and the processing sequence of writing the standard operation copy into the object identification template to obtain the object identification text can also be executed first. Alternatively, a parallel processing method can be adopted to execute the two processing processes simultaneously using multiple threads.

[0029] In actual applications, on the basis that the copy processing template includes the task text for key content extraction and the task text for resource recommendation extraction, in order to improve the accuracy of the subsequent large language model in copy processing of the copy processing text, a copy processing example can also be set in the copy processing template. In the subsequent process of inputting the copy processing text into the large language model for copy processing, the large language model can refer to the copy processing example for more accurate copy processing; similarly, the object recognition template can also include an object recognition example for object recognition. In the subsequent process of inputting the object recognition text into the large language model for object recognition, the large language model can refer to the object recognition example for more accurate object recognition.

[0030] The following uses the application of equity resource projects as an example to explain the copywriting text and its generation process, and the object recognition text and its generation process: (1) The task text in the copywriting template is: “Input a piece of text. As a resource management expert, you need to extract the key points from the text and mark them with different colors. The output is [{text, type}], where text is the original text and type is the color mark. The key contents are divided into two categories: 1. Text that indicates appeal is marked in red 2. Other key points, such as titles / subtitles, main sentences / conclusion sentences, and positive and negative content, are marked in bold. "; The copy processing text (prompt) after writing the standard service copy is: “Input a piece of text. As a resource management expert, you need to extract the key points from the text and mark them with different colors. The output is [{text, type}], where text is the original text and type is the color mark. The key contents are divided into two categories: 1. Text that indicates appeal is marked in red 2. Other key points, such as titles / subtitles, main sentences / conclusion sentences, and positive and negative content, are marked in bold.

[0031] ###example### The text entered is: "xxxxxxxx……xxxxxxxx" Output: "xxxxxxxx" Please refer to the above example to process and output the following text; The input text is: "If you can accept a six-month closed period, it is recommended that you consider the xxxx equity resource project. The project was established on xx / xx / xxxx. Since its establishment, the annualized rate of return has been 3.07%, the return in the past year has been 2.45%, the maximum drawdown is 0.45%, and the maximum drawdown repair time is 10 days. Since its establishment, the quarterly reports show that the project's rate of return has been positive in the past 6 months."".

[0032] (2) The task text in the object recognition template is: “Enter a text and extract the names of the equity projects mentioned in the text; Output a json array ["name 1", "name 2"] Requirement: If the text does not contain the name of the equity project, an empty array is returned by default; ###example### The text entered is: "xxxxxxxx……xxxxxxxx" Output: "xxxxxxxx" Please refer to the above example to process and output the following text; The input text is "If you can accept a six-month closed period, it is recommended that you consider the xxxx equity resource project. The project was established on xx / xx / xxxx. Since its establishment, the annualized rate of return has been 3.07%, the return in the past year has been 2.45%, the maximum drawdown is 0.45%, and the maximum drawdown repair time is 10 days. The quarterly reports since its establishment show that the project's rate of return has been positive in the past 6 months."".

[0033] Step S206: input the document processing text into a large language model for document processing and structural processing to obtain structured data, and input the object recognition text into a large language model for resource object recognition to obtain resource objects.

[0034] After writing the standard service copy into the copy processing template and the object recognition template to obtain the copy processing text and the object recognition text, here, the copy processing text and the object recognition text are respectively input into the large language model for corresponding processing. Specifically, on the one hand, the copy processing text is input into the large language model for copy processing and structuring processing to obtain structured data, and on the other hand, the object recognition text is input into the large language model for resource object recognition to obtain resource objects. By inputting the copy processing text and the object recognition text separately into the large language model for processing, the processing difficulty of the large language model can be reduced, thereby helping to improve the accuracy of the large language model processing, that is, improving the accuracy of the structured data and resource objects obtained by the large speech model.

[0035] In the specific implementation, when the copy processing text is input into the large language model for copy processing and structural processing, on the basis of the task text for copy processing included in the copy processing template, after the copy processing text including the copy processing template and the standard service copy is input into the large language model, the large language model performs copy processing and structural processing on the standard service copy according to the task text in the copy processing template to obtain structured data.

[0036] Specifically, in an optional implementation manner provided by this embodiment, the text processing and structured processing include: Perform semantic recognition on the task text, and extract key content from the standard operation text to obtain key content fields and / or perform resource recommendation extraction to obtain resource recommendation fields based on the semantic recognition results; Generate fields according to the parameters included in the copy processing template, and generate visualization parameters of key content fields and / or visualization parameters of resource recommendation fields.

[0037] The structured data refers to data represented by a specific data structure, such as json data represented by JavaScript Object Notation (JSON). Optionally, the structured data includes key content fields and visualization parameters of the key content fields, and / or resource recommendation fields and visualization parameters of the resource recommendation fields.

[0038] Continuing with the above example, after the above copy processing text (prompt) is input into the large language model for copy processing and structural processing, the output json data is: [{"text":"It is recommended that you consider xxxx equity resource project","type":"red"}]; Among them, "It is recommended that you consider xxxx equity resource project" is the resource recommendation field, and "red" is the visualization parameter of the resource recommendation field, which means that the resource recommendation field is displayed in red font when it is displayed.

[0039] In addition, during the specific implementation, when the object recognition text is input into the large language model for resource object recognition, the large language model performs object recognition on the standard service copy contained in the object recognition text based on the task text in the object recognition template contained in the input object recognition text, and obtains the resource object, wherein the resource object refers to the resource project, and the resource object output by the large language model can specifically be a resource object name or a resource project name.

[0040] In the specific execution process, after obtaining the resource object, the object data of the resource object can be further obtained from the resource object. Optionally, the object data includes a name field and / or resource details. Specifically, in an optional implementation provided by this embodiment, the object data of the resource object is generated in the following manner: the name field of the resource object is used as the interface call input, the retrieval interface is called to retrieve the resource details of the resource object, and the name field and resource details are used as the object data. The resource details include the object name, object type, transaction number and / or object ID (Identity document).

[0041] It should be pointed out that object data can also be structured data in a specific data structure form, that is, structured object data, which can be assembled according to name fields and resource details to generate structured object data, such as json data assembled according to name fields and resource details; based on this, object data can be replaced with structured object data, and accordingly, object data involved in subsequent processing can all be replaced with structured object data.

[0042] In this embodiment, the large language model for text processing and structural processing and the large language model for resource object recognition can be the same large language model; in addition, the large language model for text processing and structural processing and the large language model for resource object recognition can also be different large language models, the large language model for text processing and structural processing is the first large language model, and the large language model for resource object recognition is the second large language model.

[0043] In the specific implementation process, the large language model is trained in the following manner: obtain standard service texts and input them into the pre-annotated model for data annotation, and adjust the obtained annotated data to obtain training samples; fine-tune the base model to be trained based on the training samples to obtain the large language model. Optionally, the training samples include positive samples and negative samples.

[0044] Step S208: Send the structured data and the object data of the resource object to the user terminal, so as to convert the structured data and the object data into rich text tags and display them in rich text.

[0045] The above obtains structured data and resource objects through a large language model, and obtains the object data of the resource objects. On this basis, the structured data and the object data of the resource objects are sent to the user terminal of the user, so as to convert the structured data and the object data into rich text tags and display them in rich text, so as to display the content of the service provider's interactive response to the user in a rich text display manner. In the specific execution process, after sending the structured data and the object data of the resource objects to the user terminal, the user terminal converts the structured data and the object data into rich text tags and displays them in rich text.

[0046] In an optional implementation manner provided by this embodiment, the structured data and the object data are converted into rich text tags and displayed in rich text, including: Generate corresponding rich text tags based on structured data and object data, and update corresponding text fields in standard service text based on rich text tags to obtain rich text; visualize the rich text in the interactive interface with the service provider; Alternatively, corresponding rich text tags are generated according to the structured data and the object data, and corresponding text fields in the standard service text are updated based on the rich text tags corresponding to the structured data; the rich text tags corresponding to the object data are written into the updated standard service text to obtain rich text.

[0047] Specifically, in the process of generating corresponding rich text tags based on structured data and object data, on the one hand, the label parameters of the structured data mapping can be determined and the rich text tags can be generated according to the determined label parameters; on the other hand, the label parameters of the resource object mapping can be determined and the rich text tags can be generated according to the mapped label parameters.

[0048] As described above, based on the fact that the structured data includes key content fields and visualization parameters of key content fields and / or resource recommendation fields and visualization parameters of resource recommendation fields, and based on the fact that the object data includes name fields and / or resource details, in the specific implementation process, the key content fields and visualization parameters of key content fields and / or resource recommendation fields and visualization parameters of resource recommendation fields and the name fields and / or resource details may form a variety of combinations. The following provides an implementation process for generating rich text tags under three combinations. For the process of generating rich text tags under other combinations, please refer to the implementation process for generating rich text tags under the three combinations provided below. This embodiment will not be described one by one here.

[0049] In an optional implementation manner provided by this embodiment, generating corresponding rich text tags according to structured data and object data includes: Determine the label parameters of the key content fields and the visualization parameter mapping according to the preset conversion protocol, and generate corresponding rich text labels according to the key content fields and the mapped label parameters; Determine the label parameters of the name field mapping according to the preset conversion protocol, and generate the corresponding rich text label according to the name field and the mapped label parameters, or generate the corresponding rich text label according to the mapped label parameters.

[0050] In another optional implementation manner provided by this embodiment, generating corresponding rich text tags according to structured data and object data includes: Determine the label parameters of the recommended content field and the visualization parameter mapping according to the preset conversion protocol, and generate corresponding rich text labels according to the recommended content field and the mapped label parameters; Determine the label parameters of the resource details mapping according to the preset conversion protocol, and generate the corresponding rich text label according to the resource details and the mapped label parameters, or generate the corresponding rich text label according to the mapped label parameters; In a third optional implementation manner provided by this embodiment, generating corresponding rich text tags according to structured data and object data includes: Determine the label parameters of the recommended content field and the visualization parameter mapping according to the preset conversion protocol, and generate corresponding rich text labels according to the recommended content field and the mapped label parameters; Determine label parameters of the name field mapping according to a preset conversion protocol, generate object parameters of the resource detail mapping according to the visualization type corresponding to the resource detail, and generate a rich text label according to the name field and / or resource detail and the label parameters of their respective mappings.

[0051] Among them, the preset conversion protocol can be a conversion protocol pre-set for converting structured data into label parameters of rich text tags, or it can be a conversion protocol for converting the name field and / or resource details of a resource object into label parameters of rich text tags.

[0052] Optionally, the visualization type corresponding to the resource details is determined according to the object type of the resource object selected by the service provider, or according to the object type matched by the resource access record of the user in the resource service.

[0053] Using the above example, after the large language model performs text processing and structural processing, the output JSON data is: [{"text":"It is recommended that you consider xxxx equity resource project", "type":"red"}]; among them, "It is recommended that you consider xxxx equity resource project" is the resource recommendation field, and "red" is the visualization parameter of the resource recommendation field; Based on this, according to the preset conversion protocol, the resource recommendation field "It is recommended that you consider xxxx equity resource project" is mapped with the label parameter "font-weight: bold", and the visualization parameter "red" is mapped with the label parameter "color: #F93A4A". According to the resource recommendation field "It is recommended that you consider xxxx equity resource project", the mapped label parameters "font-weight: bold" and "color: #F93A4A" generate the corresponding rich text label as " It is recommended that you consider the xxxx equity resource project ”; After obtaining the rich text label, the corresponding text field in the standard service copy can be updated based on the rich text label. The updated rich text is "If you can accept the six-month closure period, It is recommended that you consider the xxxx equity resource project , the project was established on xx / xx / xxxx, with an annualized rate of return of 3.07% since its establishment, a rate of return of 2.45% in the past year, a maximum drawdown of 0.45%, and a maximum drawdown repair time of 10 days. The quarterly reports since its establishment show that the rate of return of the project has been positive in the past 6 months. "; On the other hand, after the large language model performs object recognition, the output resource object is "xxxx equity resource project", and the json data of the resource object obtained is: "{"prodNameMap":{"xxxx project":" xxxx equity resource project"}, "fundList":[{"fund_name":"xxxx equity resource project","fund_type":"equity resource","fund_code":"xxxxxx","product_id":"2022xxxxxxxxxxxx"}]}"; Among them, "prodNameMap" is the name of the resource object in the standard service copy and the corresponding full name, fundList is the specific information of the resource object, "fund_name" is the full name of the resource object, "fund_type" is the type of the resource object, "fund_code" is the transaction number, and product_id is the object ID; The generated rich text tag is: " xxxxxxxxxxxx ", and add the rich text tag to the rich text obtained after the above update; After that, the interactive interface between the user and the service personnel displayed on the user terminal is visualized to display the obtained rich text, and the display effect is as follows: Figure 3 shown.

[0054] In actual applications, in order to improve the accuracy of rich text display on the user's user terminal and reduce the possibility of display abnormalities, before the corresponding rich text label operation is generated according to the structured data and object data, the structured data and object data can also be detected. Specifically, by detecting whether the structured data and object data contain the content in the standard service copy, it is detected whether the current structured data and object data are valid. Specifically, in an optional implementation provided by this embodiment, before the corresponding rich text label operation is generated according to the structured data and object data, it also includes: detecting whether the key content field contained in the structured data and the name field contained in the object data exist in the standard service copy; if so, executing the corresponding rich text label operation according to the structured data and object data; if not, no processing is performed, or the standard service copy is displayed, specifically, the standard service copy is displayed in the interactive interface between the user and the service provider.

[0055] It should be pointed out that sending structured data and object data of resource objects to user terminals to convert the structured data and object data into rich text tags and display them in rich text can also be replaced by sending structured data and resource objects to user terminals to convert the structured data and resource objects into rich text tags and display them in rich text; specifically, the specific implementation process of converting structured data and resource objects into rich text tags and displaying them in rich text is similar to the implementation process of converting structured data and object data into rich text tags and displaying them in rich text provided above. Refer to the implementation process of converting structured data and object data into rich text tags and displaying them in rich text provided above, and replace the object data in the above implementation process with resource objects.

[0056] In summary, the data processing method based on the large language model provided in this embodiment is to first obtain the standard service copy for the service provider to interact with the user during the process of interacting with the service provider when the user accesses the resource service, and then write the standard service copy into the copy processing template and the object recognition template to obtain the copy processing text and the object recognition text, and then input the copy processing text into the large language model for copy processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects, so that by inputting the copy processing text and the object recognition text into the large language model separately for processing, the processing difficulty of the large language model can be reduced, thereby helping to improve the accuracy of the large language model processing, that is, improving the accuracy of the structured data and resource objects obtained by the large voice model; further, after obtaining the structured data and the object data of the resource object, the structured data and the object data of the resource object are sent to the user terminal of the user, so as to convert the structured data and the object data into rich text tags and display them in rich text, so as to display the content of the service provider's interactive response to the user in the form of rich text display, The above steps S202 to S208 provided in this embodiment can be executed by the server. It should be noted that the above steps S202 to S208 executed by the server and the steps S602 to S606 executed by the user terminal in the following embodiment can cooperate with each other during the execution process. Therefore, when reading this embodiment, please refer to the corresponding contents of steps S602 to S606 provided in the following method embodiment, and when reading the following method embodiment, please refer to the corresponding contents of steps S202 to S208 provided in this embodiment.

[0057] The following uses the application of a data processing method based on a large language model provided in this embodiment in the first resource service scenario as an example. Figure 4 , the data processing method based on the large language model provided in this embodiment is further described, see Figure 4 , a data processing method based on a large language model applied to a first resource service scenario specifically includes the following steps.

[0058] Step S402: obtaining a standard service document for the resource service personnel to interactively respond to the user.

[0059] Step S404, write the standard service copy into the copy processing template to obtain the copy processing text, and write the standard operation copy into the object recognition template to obtain the object recognition text.

[0060] Step S406: Input the document processing text into the large language model for document processing and structural processing to obtain first JSON data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects.

[0061] Step S408, calling the retrieval interface to retrieve the resource details of the resource object, and assembling and generating the second JSON data according to the name field and the resource details.

[0062] Step S410: Send the first JSON data and the second JSON data to the user terminal.

[0063] It should be noted that any one of steps S402 to S410 or any combination of multiple steps can be combined with any one of steps S202 to S208 to form a new implementation method according to the needs of implementation deployment; in addition, according to the needs of actual deployment, any one or multiple technical features can be selected in steps S402 to S410 and combined with any one or multiple technical features provided in steps S202 to S208 to form a new implementation method; or, any one or multiple technical features in steps S402 to S410 can be replaced with any one or multiple technical features provided in steps S202 to S208 to form a new implementation method according to the needs of actual deployment, which will not be repeated here.

[0064] In addition, it should be noted that the above steps S402 to S410 provided in this embodiment can be executed by the server. It should be noted that the above steps S402 to S410 executed by the server and the steps S412 to S420 executed by the user terminal in the following embodiment can cooperate with each other during the execution process. Therefore, when reading this embodiment, please refer to the corresponding contents of steps S412 to S420 provided in the following method embodiment, and when reading the following method embodiment, please refer to the corresponding contents of steps S402 to S410 provided in this embodiment.

[0065] The following takes the application of a data processing method based on a large language model provided in this embodiment in the second resource service scenario as an example, combined with Figure 5 , the data processing method based on the large language model provided in this embodiment is further described, see Figure 5 , a data processing method based on a large language model applied to a second resource service scenario specifically includes the following steps.

[0066] Step S502: obtaining a standard service document for interactively responding to a user by a service provider of a resource service.

[0067] Step S504, writing the standard service copy into the copy processing template and the object recognition template respectively to obtain the copy processing text and the object recognition text.

[0068] Step S506: input the document processing text into the large language model for document processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects.

[0069] Step S508: sending structured data and resource objects to the user terminal of the user.

[0070] Afterwards, the user terminal converts the structured data and resource objects into rich text tags and displays them in rich text.

[0071] In addition, sending structured data and resource objects to the user's user terminal can also be replaced by sending structured object data of structured data and resource objects to the user's user terminal; accordingly, the user terminal converts the structured data and resource objects into rich text tags and displays them in rich text, which can also be replaced by the user terminal converting the structured data and structured object data into rich text tags and displaying them in rich text.

[0072] It should be noted that any one of steps S502 to S508 or any combination of multiple steps can be combined with any one of steps S202 to S208 to form a new implementation method according to the needs of implementation deployment; in addition, according to the needs of actual deployment, any one or multiple technical features can be selected from steps S502 to S508 and combined with any one or multiple technical features provided by steps S202 to S208 to form a new implementation method; or, any one or multiple technical features in steps S502 to S508 can be replaced with any one or multiple technical features provided by steps S202 to S208 to form a new implementation method according to the needs of actual deployment, which will not be repeated here.

[0073] One or more embodiments of a data processing method based on a large language model provided in this specification are as follows: Reference Figure 6 The data processing method based on the large language model provided in this embodiment specifically includes steps S602 to S606.

[0074] Step S602: obtaining a standard service document, structured data and object data of a resource object for interactively responding to a user by a service provider of the resource service.

[0075] The resource service described in this embodiment refers to the resource-related services provided within the application, such as the resource services provided by the application for recommending resource items, trading, and other related processing, where the resource items can be resource items for resource management, and the managed resources can be equity resources, such as asset types that can be used for trading (funds, stocks, bonds), etc. The managed resources can also be financial resources or virtual resources, and the virtual resources can be virtual resources such as carbon savings and points.

[0076] The service provider of a resource service refers to the entity assigned to the user to provide services during the access process of the resource service. The service provider may be the service personnel of the service provider, such as customer service personnel or professional service personnel. The service provider may also be a virtual customer service. The service provider may be assigned to the user by the service platform of the resource service. The service platform may be the operator, provider or developer of the application. It may also be assigned to the user by the resource provider or resource transaction party of the resource service. Specifically, the service provider may be assigned to the user if the user data and / or resource data of the user meets preset conditions. For example, a service personnel may be assigned to the user if the user's resource access data and / or resource transaction data of the resource service meets preset conditions. For another example, a service personnel may be assigned to the user if the user's level in the application is higher than the preset level.

[0077] In this embodiment, in order to improve the service experience of users in the process of accessing resource services, in the process of interaction between the service provider of resource services and users, a standardized service process (standard service process) can be set for the service provider, and a standardized service document, i.e., a standard service document, can be set for each service node in the standard service process; in the specific implementation process, in the scenario where the service provider of resource services interacts with users, in the process of determining the standard service document for the service provider to respond to the user's interaction, a certain service node (target node) standard service document can be selected in the service document according to the service document of the standard service process as the corresponding standard service document for interacting with the user; In addition, the standard service copy can be determined by matching the user's resource access record with the standard service process of the resource service, that is, determining the standard service copy that matches the resource access record in the standard service process of the resource service; or, the standard service copy can be determined based on the target node selected by the service provider in the standard service process of the resource service, that is, using the standard service copy of the target node selected by the service provider in the standard service process of the resource service as the standard service copy for the service provider to interact with the user.

[0078] Optionally, the service party interactively responds to the user after the service party triggers the target node in the standard service process of the resource service, or after matching the user's resource access record with the standard service process of the resource service to obtain the target node.

[0079] Correspondingly, the standard service copy can be determined after matching the user's resource access record with the standard service process of the resource service, or it can also be determined after the service provider triggers the target node in the standard service process of the resource service, that is: determine the standard service copy based on the target node triggered by the service provider.

[0080] For example, when a user interacts with the service provider during the process of accessing a resource service, the service provider selects the target node in the standard service process of the resource service. The standard service copy based on the target node is "If you can accept a six-month closed period, it is recommended that you consider the xxxx equity resource project. The project was established on xx / xx / xxxx. Since its establishment, the annualized rate of return has been 3.07%, the rate of return in the past year has been 2.45%, the maximum drawdown is 0.45%, and the maximum drawdown repair time is 10 days. The quarterly reports since its establishment show that the project's rate of return has been positive for the past 6 months."

[0081] The structured data described in this embodiment refers to data represented by a specific data structure, for example, the structured data may be json data represented in the form of json (JavaScript Object Notation). Optionally, the structured data includes key content fields and visualization parameters of the key content fields, and / or resource recommendation fields and visualization parameters of the resource recommendation fields.

[0082] Optionally, structured data is obtained by calling a large language model for text processing and structural processing.

[0083] In the specific execution process, the structured data is obtained by inputting the copy processing text into the large language model for copy processing and structural processing. In the process of inputting the copy processing text into the large language model for copy processing and structural processing, on the basis of including the task text for copy processing in the copy processing template, after the copy processing text including the copy processing template and the standard service copy is input into the large language model, the large language model performs copy processing and structural processing on the standard service copy according to the task text in the copy processing template to obtain structured data. Among them, the process of obtaining the copy processing text refers to the process of obtaining the copy processing text provided in step S204 in the above embodiment.

[0084] Specifically, in an optional implementation manner provided by this embodiment, the text processing and structured processing include: Perform semantic recognition on the task text, and extract key content from the standard operation text to obtain key content fields and / or perform resource recommendation extraction to obtain resource recommendation fields based on the semantic recognition results; Generate fields according to the parameters included in the copy processing template, and generate visualization parameters of key content fields and / or visualization parameters of resource recommendation fields.

[0085] Continuing with the above example, after the above copy processing text (prompt) is input into the large language model for copy processing and structural processing, the output json data is: [{"text":"It is recommended that you consider xxxx equity resource project","type":"red"}]; Among them, "It is recommended that you consider xxxx equity resource project" is the resource recommendation field, and "red" is the visualization parameter of the resource recommendation field, which means that the resource recommendation field is displayed in red font when it is displayed.

[0086] Optionally, the resource object is obtained by calling a large language model to perform resource object recognition. In the specific execution process, when the object recognition text is input into the large language model for resource object recognition, the large language model performs object recognition on the standard service copy contained in the object recognition text according to the task text in the object recognition template contained in the input object recognition text, and obtains the resource object, wherein the resource object refers to the resource project, and the resource object output by the large language model can specifically be the resource object name or the resource project name. For the process of obtaining the object recognition text, refer to the process of obtaining the object recognition text provided in step S204 in the above embodiment.

[0087] After obtaining the resource object, the object data of the resource object is further obtained from the resource object. Optionally, the object data includes a name field and / or resource details. Specifically, in an optional implementation provided by this embodiment, the object data of the resource object is generated in the following manner: the name field of the resource object is used as an interface call input, the retrieval interface is called to retrieve the resource details of the resource object, and the name field and resource details are used as object data. Resource details include object name, object type, transaction number and / or object ID (Identity document).

[0088] It should be pointed out that object data can also be structured data in a specific data structure form, that is, structured object data, which can be assembled according to name fields and resource details to generate structured object data, such as json data assembled according to name fields and resource details; based on this, object data can be replaced with structured object data, and accordingly, object data involved in subsequent processing can all be replaced with structured object data.

[0089] In this embodiment, the large language model for text processing and structural processing and the large language model for resource object recognition can be the same large language model; in addition, the large language model for text processing and structural processing and the large language model for resource object recognition can also be different large language models, the large language model for text processing and structural processing is the first large language model, and the large language model for resource object recognition is the second large language model.

[0090] In the specific implementation process, the large language model is trained in the following manner: obtain standard service texts and input them into the pre-annotated model for data annotation, and adjust the obtained annotated data to obtain training samples; fine-tune the base model to be trained based on the training samples to obtain the large language model. Optionally, the training samples include positive samples and negative samples.

[0091] Step S604: Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text.

[0092] After obtaining the structured data and the object data of the resource object, the corresponding rich text tags are generated according to the structured data and the object data, and the standard service copy is updated based on the generated rich text tags to obtain the rich text. Specifically, the corresponding copy fields in the standard service copy are updated based on the generated rich text tags to obtain the rich text.

[0093] In an optional implementation manner provided by this embodiment, the structured data and the object data are converted into rich text tags and displayed in rich text, including: Generate corresponding rich text tags based on structured data and object data, and update corresponding text fields in standard service text based on rich text tags to obtain rich text; visualize the rich text in the interactive interface with the service provider; Alternatively, corresponding rich text tags are generated according to the structured data and the object data, and corresponding text fields in the standard service text are updated based on the rich text tags corresponding to the structured data; the rich text tags corresponding to the object data are written into the updated standard service text to obtain rich text.

[0094] Specifically, in the process of generating corresponding rich text tags based on structured data and object data, on the one hand, the label parameters of the structured data mapping can be determined and the rich text tags can be generated according to the determined label parameters; on the other hand, the label parameters of the resource object mapping can be determined and the rich text tags can be generated according to the mapped label parameters.

[0095] As described above, based on the fact that the structured data includes key content fields and visualization parameters of key content fields and / or resource recommendation fields and visualization parameters of resource recommendation fields, and based on the fact that the object data includes name fields and / or resource details, in the specific implementation process, the key content fields and visualization parameters of key content fields and / or resource recommendation fields and visualization parameters of resource recommendation fields and the name fields and / or resource details may form a variety of combinations. The following provides an implementation process for generating rich text tags under three combinations. For the process of generating rich text tags under other combinations, please refer to the implementation process for generating rich text tags under the three combinations provided below. This embodiment will not be described one by one here.

[0096] In an optional implementation manner provided by this embodiment, generating corresponding rich text tags according to structured data and object data includes: Determine the label parameters of the key content fields and the visualization parameter mapping according to the preset conversion protocol, and generate corresponding rich text labels according to the key content fields and the mapped label parameters; Determine the label parameters of the name field mapping according to the preset conversion protocol, and generate the corresponding rich text label according to the name field and the mapped label parameters, or generate the corresponding rich text label according to the mapped label parameters.

[0097] In another optional implementation manner provided by this embodiment, generating corresponding rich text tags according to structured data and object data includes: Determine the label parameters of the recommended content field and the visualization parameter mapping according to the preset conversion protocol, and generate corresponding rich text labels according to the recommended content field and the mapped label parameters; Determine the label parameters of the resource details mapping according to the preset conversion protocol, and generate the corresponding rich text label according to the resource details and the mapped label parameters, or generate the corresponding rich text label according to the mapped label parameters; In a third optional implementation manner provided by this embodiment, generating corresponding rich text tags according to structured data and object data includes: Determine the label parameters of the recommended content field and the visualization parameter mapping according to the preset conversion protocol, and generate corresponding rich text labels according to the recommended content field and the mapped label parameters; Determine label parameters of the name field mapping according to a preset conversion protocol, generate object parameters of the resource detail mapping according to the visualization type corresponding to the resource detail, and generate a rich text label according to the name field and / or resource detail and the label parameters of their respective mappings.

[0098] Among them, the preset conversion protocol can be a conversion protocol pre-set for converting structured data into label parameters of rich text tags, or it can be a conversion protocol for converting the name field and / or resource details of a resource object into label parameters of rich text tags.

[0099] Optionally, the visualization type corresponding to the resource details is determined according to the object type of the resource object selected by the service provider, or according to the object type matched by the resource access record of the user in the resource service.

[0100] Using the above example, after the large language model performs text processing and structural processing, the output JSON data is: [{"text":"It is recommended that you consider xxxx equity resource project", "type":"red"}]; among them, "It is recommended that you consider xxxx equity resource project" is the resource recommendation field, and "red" is the visualization parameter of the resource recommendation field; Based on this, according to the preset conversion protocol, the resource recommendation field "It is recommended that you consider xxxx equity resource project" is mapped with the label parameter "font-weight: bold", and the visualization parameter "red" is mapped with the label parameter "color: #F93A4A". According to the resource recommendation field "It is recommended that you consider xxxx equity resource project", the mapped label parameters "font-weight: bold" and "color: #F93A4A" generate the corresponding rich text label as " It is recommended that you consider the xxxx equity resource project ”; After obtaining the rich text label, the corresponding text field in the standard service copy can be updated based on the rich text label. The updated rich text is "If you can accept the six-month closure period, It is recommended that you consider the xxxx equity resource project , the project was established on xx / xx / xxxx, with an annualized rate of return of 3.07% since its establishment, a rate of return of 2.45% in the past year, a maximum drawdown of 0.45%, and a maximum drawdown repair time of 10 days. The quarterly reports since its establishment show that the rate of return of the project has been positive in the past 6 months. "; On the other hand, after the large language model performs object recognition, the output resource object is "xxxx equity resource project", and the json data of the resource object obtained is: "{"prodNameMap":{"xxxx project":" xxxx equity resource project"}, "fundList":[{"fund_name":"xxxx equity resource project","fund_type":"equity resource","fund_code":"xxxxxx","product_id":"2022xxxxxxxxxxxx"}]}"; Among them, "prodNameMap" is the name of the resource object in the standard service copy and the corresponding full name, fundList is the specific information of the resource object, "fund_name" is the full name of the resource object, "fund_type" is the type of the resource object, "fund_code" is the transaction number, and product_id is the object ID; The generated rich text tag is: " xxxxxxxxxxxx ", and add the rich text tag to the rich text obtained after the above update; In actual applications, in order to improve the accuracy of rich text display on the user's user terminal and reduce the possibility of display abnormalities, before the corresponding rich text label operation is generated according to the structured data and object data, the structured data and object data can also be detected. Specifically, by detecting whether the structured data and object data contain the content in the standard service copy, it is detected whether the current structured data and object data are valid. Specifically, in an optional implementation provided by this embodiment, before the corresponding rich text label operation is generated according to the structured data and object data, it also includes: detecting whether the key content field contained in the structured data and the name field contained in the object data exist in the standard service copy; if so, executing the corresponding rich text label operation according to the structured data and object data; if not, not doing it, or displaying the standard service copy, specifically displaying the standard service copy in the interactive interface between the user and the service provider.

[0101] Step S606: Visually display the rich text in the interactive interface with the service provider.

[0102] In specific implementation, based on the rich text obtained by updating the corresponding text field in the standard service text based on the rich text tag, the rich text is visually displayed, specifically, the rich text is visually displayed in the interactive interface between the user and the service provider, so as to display the content of the service provider's interactive response to the user in the form of rich text display. For example, the rich text obtained above is visually displayed in the interactive interface between the user and the service personnel displayed on the user terminal, and the display effect is as follows: Figure 3 shown.

[0103] The above steps S602 to S606 provided in this embodiment can be executed by the user terminal. It should be noted that the above steps S202 to S208 executed by the user terminal and the steps S202 to S208 executed by the server in the above embodiment can cooperate with each other during the execution process. Therefore, when reading this embodiment, please refer to the corresponding contents of steps S202 to S208 provided in the above method embodiment, and when reading the above method embodiment, please refer to the corresponding contents of steps S602 to S606 provided in this embodiment.

[0104] The following uses the data processing method based on a large language model provided in this embodiment as an example of the application in the first live interactive scene. Figure 4 , the data processing method based on the large language model provided in this embodiment is further described, see Figure 4 , a data processing method based on a large language model applied to the first live interactive scenario specifically includes the following steps.

[0105] Step S412, receiving the standard service text of the resource service, the first JSON data and the second JSON data sent by the server.

[0106] Optionally, the first JSON data is obtained by calling a large language model for text processing and structural processing, and the resource object is obtained by calling a large language model for resource object identification; the second JSON data is generated by assembling the name field and resource details of the resource object.

[0107] Step S414, detecting whether the key content field included in the first json data and the name field included in the second json data exist in the standard service document; If yes, execute the following steps S416 to S420; If not, no action is taken, or standard service text is displayed, specifically, standard service text is displayed in the interaction interface between the user and the service provider.

[0108] Step S416: Generate corresponding rich text tags according to the first JSON data and the second JSON data.

[0109] Step S418: Update the corresponding text field in the standard service text based on the generated rich text tag to obtain rich text.

[0110] Step S420: Visually display the rich text in the interactive interface between the user and the service provider.

[0111] It should be noted that any one of steps S412 to S420 or any combination of multiple steps can be combined with any one of steps S602 to S606 to form a new implementation method according to the needs of implementation deployment; in addition, according to the needs of actual deployment, any one or multiple technical features can be selected from steps S412 to S420 and combined with any one or multiple technical features provided by steps S602 to S606 to form a new implementation method; or, any one or multiple technical features in steps S412 to S420 can be replaced with any one or multiple technical features provided by steps S602 to S606 to form a new implementation method according to the needs of actual deployment, which will not be elaborated here.

[0112] An embodiment of a data processing device based on a large language model provided in this specification is as follows: In the above-mentioned embodiment, a data processing method based on a large language model is provided, and correspondingly, a data processing device based on a large language model is also provided, which is described below with reference to the accompanying drawings.

[0113] Reference Figure 7 , which shows a schematic diagram of an embodiment of a data processing device based on a large language model provided in this embodiment.

[0114] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the relevant parts can refer to the corresponding description of the method embodiment provided above. The device embodiment described below is only illustrative.

[0115] This embodiment provides a data processing device based on a large language model, the device comprising: The document acquisition module 702 is configured to acquire a standard service document for interactive response of a service provider of a resource service to a user; The text generation module 704 is configured to write the standard service text into the text processing template to obtain the text processing text, and write the standard operation text into the object recognition template to obtain the object recognition text; The text processing module 706 is configured to input the text processing text into the large language model for text processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects; The data sending module 708 is configured to send the structured data and the object data of the resource object to the user terminal, so as to convert the structured data and the object data into rich text tags and display them in rich text.

[0116] Another embodiment of a data processing device based on a large language model provided in this specification is as follows: In the above-mentioned embodiment, another data processing method based on a large language model is provided. Correspondingly, another data processing device based on a large language model is also provided, which will be described below with reference to the accompanying drawings.

[0117] Reference Figure 8 , which shows a schematic diagram of an embodiment of a data processing device based on a large language model provided in this embodiment.

[0118] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the relevant parts can refer to the corresponding description of the method embodiment provided above. The device embodiment described below is only illustrative.

[0119] This embodiment provides a data processing device based on a large language model, the device comprising: The data receiving module 802 is configured to obtain standard service text, structured data and object data of resource objects for interactive response of the service provider of the resource service to the user; the structured data is obtained by calling the large language model to perform text processing and structure processing, and the resource objects are obtained by calling the large language model to perform resource object recognition; The tag generation module 804 is configured to generate corresponding rich text tags according to the structured data and the object data, and update the corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text display module 806 is configured to perform a visual display of the rich text in the interactive interface with the service provider.

[0120] An embodiment of a data processing device based on a large language model provided in this specification is as follows: Corresponding to the data processing method based on a large language model described above, based on the same technical concept, one or more embodiments of this specification further provide a data processing device based on a large language model, the data processing device based on a large language model is used to execute the data processing method based on a large language model provided above, Fig. 9 A schematic diagram of the structure of a data processing device based on a large language model provided for one or more embodiments of this specification.

[0121] This embodiment provides a data processing device based on a large language model, including: like Fig. 9As shown, the data processing device based on the large language model may have relatively large differences due to different configurations or performances, and may include one or more processors 901 and memory 902, and the memory 902 may store one or more storage applications or data. Among them, the memory 902 may be a short-term storage or a persistent storage. The application stored in the memory 902 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the data processing device based on the large language model. Furthermore, the processor 901 may be configured to communicate with the memory 902, and execute a series of computer executable instructions in the memory 902 on the data processing device based on the large language model. The data processing device based on the large language model may also include one or more power supplies 903, one or more wired or wireless network interfaces 904, one or more input / output interfaces 905, one or more keyboards 906, etc.

[0122] In a specific embodiment, a data processing device based on a large language model includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions for the data processing device based on the large language model, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following: Obtain the standard service copy of the resource service provider's interactive response to the user; Writing the standard service document into a document processing template to obtain a document processing text, and writing the standard operation document into an object recognition template to obtain an object recognition text; Input the document processing text into the large language model for document processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects; The structured data and the object data of the resource object are sent to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

[0123] Another embodiment of a data processing device based on a large language model provided in this specification is as follows: Corresponding to the other data processing method based on a large language model described above, based on the same technical concept, one or more embodiments of this specification also provide another data processing device based on a large language model, and the data processing device based on a large language model is used to execute the other data processing method based on a large language model provided above. Fig.10A schematic diagram of the structure of another data processing device based on a large language model provided for one or more embodiments of this specification.

[0124] This embodiment provides a data processing device based on a large language model, including: like Fig.10 As shown, the data processing device based on the large language model may have relatively large differences due to different configurations or performances, and may include one or more processors 1001 and memory 1002, and the memory 1002 may store one or more storage applications or data. Among them, the memory 1002 may be a short-term storage or a persistent storage. The application stored in the memory 1002 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the data processing device based on the large language model. Furthermore, the processor 1001 may be configured to communicate with the memory 1002, and execute a series of computer executable instructions in the memory 1002 on the data processing device based on the large language model. The data processing device based on the large language model may also include one or more power supplies 1003, one or more wired or wireless network interfaces 1004, one or more input / output interfaces 1005, etc.

[0125] In a specific embodiment, a data processing device based on a large language model includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions for the data processing device based on the large language model, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following: Obtaining standard service text, structured data and object data of resource objects for interactive response from the service provider of the resource service to the user; the structured data is obtained by calling the large language model for text processing and structure processing, and the resource objects are obtained by calling the large language model for resource object recognition; Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text is visually displayed in the interactive interface with the service provider.

[0126] An embodiment of a computer-readable storage medium provided in this specification is as follows: Corresponding to the above-described data processing method based on a large language model, based on the same technical concept, one or more embodiments of this specification also provide a computer-readable storage medium.

[0127] The computer-readable storage medium provided in this embodiment is used to store computer-executable instructions, and the computer-executable instructions implement the following process when executed: Obtain the standard service copy of the resource service provider's interactive response to the user; Writing the standard service document into a document processing template to obtain a document processing text, and writing the standard operation document into an object recognition template to obtain an object recognition text; Input the document processing text into the large language model for document processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects; The structured data and the object data of the resource object are sent to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

[0128] It should be noted that an embodiment of a computer-readable storage medium in this specification and an embodiment of a data processing method based on a large language model in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0129] Another computer-readable storage medium embodiment provided in this specification is as follows: Corresponding to the other data processing method based on a large language model described above, based on the same technical concept, one or more embodiments of this specification also provide another computer-readable storage medium.

[0130] The computer-readable storage medium provided in this embodiment is used to store computer-executable instructions, and the computer-executable instructions implement the following process when executed: Obtaining standard service text, structured data and object data of resource objects for interactive response from the service provider of the resource service to the user; the structured data is obtained by calling the large language model for text processing and structure processing, and the resource objects are obtained by calling the large language model for resource object recognition; Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text is visually displayed in the interactive interface with the service provider.

[0131] It should be noted that the embodiment of another computer-readable storage medium in this specification and the embodiment of another data processing method based on a large language model in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0132] An embodiment of a computer program product provided in this specification is as follows: Corresponding to the data processing method based on a large language model described above, based on the same technical concept, one or more embodiments of this specification also provide a computer program product.

[0133] A computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the following steps: Obtain the standard service copy of the resource service provider's interactive response to the user; Writing the standard service document into a document processing template to obtain a document processing text, and writing the standard operation document into an object recognition template to obtain an object recognition text; Input the document processing text into the large language model for document processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects; The structured data and the object data of the resource object are sent to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

[0134] It should be noted that an embodiment of a computer program product in this specification and an embodiment of a data processing method based on a large language model in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0135] Another computer program product embodiment provided in this specification is as follows: Corresponding to the other data processing method based on a large language model described above, based on the same technical concept, one or more embodiments of this specification also provide another computer program product.

[0136] A computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the following steps: Obtaining standard service text, structured data and object data of resource objects for interactive response from the service provider of the resource service to the user; the structured data is obtained by calling the large language model for text processing and structure processing, and the resource objects are obtained by calling the large language model for resource object recognition; Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text is visually displayed in the interactive interface with the service provider.

[0137] It should be noted that an embodiment of another computer program product in this specification and an embodiment of another data processing method based on a large language model in this specification are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0138] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments. For example, the device embodiment, equipment embodiment, computer-readable storage medium embodiment, and computer program product embodiment are similar to the method embodiment, so the description is relatively simple. To read the relevant contents of the device embodiment, equipment embodiment, computer-readable storage medium embodiment, and computer program product embodiment, please refer to the partial description of the method embodiment.

[0139] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0140] In the 1930s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to circuit structures such as diodes, transistors, switches, etc.) or software improvements (improvements to method flows). However, with the development of technology, many of today's method flow improvements can be regarded as direct improvements to hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using hardware entity modules. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask chip manufacturers to design and produce dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one kind of HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0141] The controller may be implemented in any suitable manner, for example, the controller may take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (e.g., software or firmware) executable by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, and the memory controller may also be implemented as part of the control logic of the memory. It is also known to those skilled in the art that, in addition to implementing the controller in a purely computer-readable program code manner, the controller may be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, such a controller may be considered as a hardware component, and the devices for implementing various functions included therein may also be considered as structures within the hardware component. Or even, the devices for implementing various functions may be considered as both software modules for implementing the method and structures within the hardware component.

[0142] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0143] For the convenience of description, the above devices are described in terms of functions and are divided into various units. Of course, when implementing the embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0144] It should be understood by those skilled in the art that one or more embodiments of the present specification may be provided as a method, system or computer program product. Therefore, one or more embodiments of the present specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present specification may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable large language model-based data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable large language model-based data processing device generate instructions for implementing the processes in the flowchart. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0146] These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device based on a large language model to work in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0147] These computer program instructions can also be loaded onto a computer or other programmable data processing device based on a large language model, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0148] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0149] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0150] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0151] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "includes at least one ..." do not exclude the presence of other identical elements in the process, method, commodity or device including the elements.

[0152] One or more embodiments of the present specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0153] The above description is only an embodiment of this document and is not intended to limit this document. For those skilled in the art, this document may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this document should be included in the scope of the claims of this document.

Claims

1. A data processing method based on a large language model, comprising: Obtain the standard service copy of the resource service provider's interactive response to the user; Writing the standard service document into a document processing template to obtain a document processing text, and writing the standard operation document into an object recognition template to obtain an object recognition text; Input the document processing text into the large language model for document processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects; The structured data and the object data of the resource object are sent to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

2. According to the data processing method based on a large language model in claim 1, the converting the structured data and the object data into rich text tags and displaying them in rich text comprises: Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text is visually displayed in the interactive interface with the service provider.

3. According to the data processing method based on a large language model in claim 2, before the operation of generating corresponding rich text tags according to the structured data and the object data is performed, it also includes: Detecting whether a key content field included in the structured data and a name field included in the object data exist in the standard service document; If so, the operation of generating a corresponding rich text tag according to the structured data and the object data is performed.

4. According to the data processing method based on a large language model in claim 2, the structured data comprises key content fields and visualization parameters of the key content fields, and / or resource recommendation fields and visualization parameters of the resource recommendation fields; The object data includes a name field and / or resource details.

5. According to the data processing method based on a large language model as claimed in claim 4, the generating corresponding rich text tags according to the structured data and the object data comprises: Determine the label parameters mapped between the key content field and the visualization parameter according to a preset conversion protocol, and generate corresponding rich text labels according to the key content field and the mapped label parameters; The label parameters of the name field mapping are determined according to a preset conversion protocol, and a corresponding rich text label is generated according to the name field and the mapped label parameters.

6. According to the data processing method based on a large language model in claim 4, generating corresponding rich text tags according to the structured data and the object data comprises: Determine the label parameters mapped by the key content field and the visualization parameter according to the preset conversion protocol, and generate corresponding rich text labels according to the key content field and the mapped label parameters; Determine the label parameters of the name field mapping according to the preset conversion protocol, generate the object parameters of the resource detail mapping according to the visualization type corresponding to the resource details, and generate a rich text label according to the name field, the resource details and the label parameters of each mapping.

7. According to the data processing method based on a large language model in claim 6, the visualization type corresponding to the resource details is determined according to the object type of the resource object selected by the service provider, or determined according to the object type matched by the resource access record of the user in the resource service.

8. According to the data processing method based on a large language model as claimed in claim 1, the copywriting processing template comprises: Task text for key content extraction and / or task text for resource recommendation extraction; The document processing and structural processing include: Performing semantic recognition on the task text, and extracting key content from the standard operation text to obtain key content fields and / or performing resource recommendation extraction to obtain resource recommendation fields according to the semantic recognition result; Generate fields according to the parameters included in the document processing template, and generate visualization parameters of key content fields and / or visualization parameters of resource recommendation fields.

9. According to the data processing method based on a large language model in claim 1, the object data of the resource object is generated in the following manner: The name field of the resource object is used as an interface call input, a retrieval interface is called to retrieve resource details of the resource object, and the name field and the resource details are used as the object data.

10. According to the data processing method based on a large language model according to claim 1, the service party interactively responds to the user after the service party triggers the target node in the standard service process of the resource service, or after matching the user's resource access record with the standard service process of the resource service to obtain the target node.

11. According to the data processing method based on a large language model as described in claim 1, the large language model for copywriting processing and structural processing and the large language model for resource object recognition are the same large language model.

12. The data processing method based on a large language model according to claim 11, wherein the large language model is trained in the following manner: Obtain standard service documents and input them into the pre-annotated model for data annotation, and adjust the obtained annotated data to obtain training samples; The large language model is obtained by fine-tuning the base model to be trained based on the training samples; wherein, The training samples include positive samples and negative samples.

13. A data processing method based on a large language model, comprising: Obtain the standard service text, structured data and object data of resource objects for interactive responses from the service provider of the resource service to the user; The structured data is obtained by calling a large language model to perform text processing and structural processing, and the resource object is obtained by calling a large language model to perform resource object recognition; Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text is visually displayed in the interactive interface with the service provider.

14. According to the data processing method based on a large language model in claim 13, after the step of obtaining the standard service copy, structured data and object data of the resource object for the service provider of the resource service to interactively respond to the user is executed, and before the step of generating corresponding rich text tags according to the structured data and the object data, and updating the corresponding copy fields in the standard service copy based on the rich text tags to obtain the rich text is executed, it also includes: Detecting whether a key content field included in the structured data and a name field included in the object data exist in the standard service document; If so, the step of generating corresponding rich text tags according to the structured data and the object data, and updating corresponding text fields in the standard service text based on the rich text tags to obtain rich text is executed.

15. According to the data processing method based on a large language model of claim 13, the structured data comprises key content fields and visualization parameters of the key content fields, and / or resource recommendation fields and visualization parameters of the resource recommendation fields; The object data includes a name field and / or resource details.

16. According to the data processing method based on a large language model of claim 15, generating corresponding rich text tags according to the structured data and the object data comprises: Determine the label parameters mapped between the key content field and the visualization parameter according to a preset conversion protocol, and generate corresponding rich text labels according to the key content field and the mapped label parameters; The label parameters of the name field mapping are determined according to a preset conversion protocol, and a corresponding rich text label is generated according to the name field and the mapped label parameters.

17. According to the data processing method based on a large language model of claim 15, generating corresponding rich text tags according to the structured data and the object data comprises: Determine the label parameters mapped by the key content field and the visualization parameter according to the preset conversion protocol, and generate corresponding rich text labels according to the key content field and the mapped label parameters; Determine the label parameters of the name field mapping according to the preset conversion protocol, generate the object parameters of the resource detail mapping according to the visualization type corresponding to the resource details, and generate a rich text label according to the name field, the resource details and the label parameters of each mapping.

18. According to the data processing method based on a large language model in claim 17, the visualization type corresponding to the resource details is determined according to the object type of the resource object selected by the service provider, or determined according to the object type matched by the resource access record of the user in the resource service.

19. A data processing device based on a large language model, comprising: A document acquisition module, configured to acquire a standard service document for interactive response from a service provider of a resource service to a user; A text generation module is configured to write the standard service copy into a copy processing template to obtain a copy processing text, and write the standard operation copy into an object recognition template to obtain an object recognition text; A text processing module is configured to input the text processing text into a large language model for text processing and structural processing to obtain structured data, and input the object recognition text into a large language model for resource object recognition to obtain a resource object; The data sending module is configured to send the structured data and the object data of the resource object to the user terminal, so as to convert the structured data and the object data into rich text tags and display them in rich text.

20. A data processing device based on a large language model, comprising: A data receiving module is configured to obtain a standard service document, structured data and object data of a resource object for interactive response from a service provider of a resource service to a user; The structured data is obtained by calling a large language model to perform text processing and structural processing, and the resource object is obtained by calling a large language model to perform resource object recognition; A label generation module is configured to generate corresponding rich text labels according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text labels to obtain rich text; The rich text display module is configured to perform a visual display of the rich text in the interactive interface with the service provider.

21. A data processing device based on a large language model, comprising: processor; and a memory configured to store computer executable instructions that, when executed, cause the processor to: Obtain the standard service copy of the resource service provider's interactive response to the user; Writing the standard service document into a document processing template to obtain a document processing text, and writing the standard operation document into an object recognition template to obtain an object recognition text; Input the document processing text into the large language model for document processing and structural processing to obtain structured data, and input the object recognition text into the large language model for resource object recognition to obtain resource objects; The structured data and the object data of the resource object are sent to a user terminal to convert the structured data and the object data into rich text tags and display them in rich text.

22. A data processing device based on a large language model, comprising: processor; and a memory configured to store computer executable instructions that, when executed, cause the processor to: Obtaining standard service text, structured data and object data of resource objects for interactive response from the service provider of the resource service to the user; the structured data is obtained by calling the large language model for text processing and structure processing, and the resource objects are obtained by calling the large language model for resource object recognition; Generate corresponding rich text tags according to the structured data and the object data, and update corresponding text fields in the standard service text based on the rich text tags to obtain rich text; The rich text is visually displayed in the interactive interface with the service provider.

23. A computer-readable storage medium for storing computer-executable instructions, wherein the computer-executable instructions implement the steps of the method of claim 1 or 13 when executed.

Images