Information processing method and device based on large language model and storage medium
By using a large language model to determine and call the operation results of external services to generate reply content, the accuracy and logical coherence issues of large language models under complex input are resolved, and more efficient information processing is achieved.
Patent Information
- Application Number
- CN202510757084.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-06
- Publication Date
- 2025-10-10
AI Technical Summary
How to enable large language models to accurately generate responses when faced with complex input content, especially when involving multiple external services, while maintaining logical coherence and accuracy.
Through the large language model, the target external services and their operations involved in the input content are determined, these services are called to perform corresponding operations, and reply content is generated based on the operation results. The calling order and result integration of multiple services are optimized to improve accuracy.
Improves the accuracy and logical coherence of responses generated by large language models, ensuring consistency and accuracy when multiple external services are involved.
Smart Images
Figure CN120763281A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the field of artificial intelligence technology, in particular to technical fields such as large models, deep learning, and natural language processing, and in particular to an information processing method, device, and storage medium based on a large language model. Background Art
[0002] With the rapid development of artificial intelligence (AI), large language models have been widely used across various industries, particularly in natural language processing. As input complexity increases, ensuring that large language models accurately respond to input content becomes a pressing technical challenge. Summary of the Invention
[0003] The present disclosure provides an information processing method, device, and storage medium based on a large language model.
[0004] According to one aspect of the present disclosure, a large language model-based information processing method is provided, the method comprising: obtaining input content; determining, based on the large language model, that the input content involves a target external service and a target operation to be performed by the target external service; using the large language model, calling the target external service to perform the target operation; and using the large language model, generating reply content to the input content based on an operation result of the target external service performing the target operation.
[0005] According to another aspect of the present disclosure, an information processing device based on a large language model is provided, the device comprising: an acquisition module for acquiring input content; a first determination module for determining, based on the large language model, whether the input content involves a target external service and a target operation to be performed by the target external service; a calling module for using the large language model to call the target external service to perform the target operation; and a generation module for using the large language model to generate reply content to the input content according to an operation result of the target external service performing the target operation.
[0006] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the information processing method based on the large language model proposed in the present disclosure.
[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable a computer to execute the information processing method based on the large language model proposed in the present disclosure.
[0008] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which, when executed by a processor, implements the steps of the information processing method based on a large language model proposed in the present disclosure.
[0009] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure.
[0011] Figure 1 is a schematic diagram according to a first embodiment of the present disclosure;
[0012] Figure 2 is a schematic diagram according to a second embodiment of the present disclosure;
[0013] Figure 3 is a schematic diagram according to a third embodiment of the present disclosure;
[0014] Figure 4 is a schematic diagram according to a fourth embodiment of the present disclosure;
[0015] Figure 5 is a schematic diagram according to the fifth embodiment of the present disclosure
[0016] Figure 6 It is a block diagram of an electronic device used to implement the information processing method based on a large language model according to an embodiment of the present disclosure. DETAILED DESCRIPTION
[0017] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.
[0018] Figure 1 This is a schematic diagram according to the first embodiment of the present disclosure. It should be noted that the information processing method based on the large language model of the embodiment of the present disclosure can be applied to an information processing device based on the large language model, which can be an electronic device, or can be configured in an electronic device.
[0019] Among them, the electronic device can be any device with computing capabilities, such as a personal computer (PC), a mobile terminal, a server, etc. The mobile terminal can be, for example, a vehicle-mounted device, a mobile phone, a tablet computer, a personal digital assistant, a wearable device, a smart speaker, a server, a server cluster, and other hardware devices with various operating systems, touch screens and / or display screens.
[0020] It should be noted that, in the following embodiments, an information processing device based on a large language model is described as an electronic device as an example.
[0021] like Figure 1 As shown, the information processing method based on the large language model may include the following steps:
[0022] Step 101: Obtain input content.
[0023] The input content may be input by a user, for example, the input content may be input by a user through natural language.
[0024] Step 102: Determine, based on the large language model, whether the input content involves a target external service and a target operation that the target external service needs to perform.
[0025] In this embodiment, the input content may be input into a large language model, so as to determine through the large language model whether the input content relates to a target external service and a target operation that needs to be performed by the target external service.
[0026] In some embodiments, a large language model is used to identify the intent of the input content, obtaining multiple operational intents associated with the input content. External resources and external services whose functional descriptions match the operational intents are selected as target external resources and external services from a list of external resources and external services. The target operation to be performed by the external service is then determined based on the large language model. This allows the large language model to accurately determine the target external service involved in the input content, based on the operational intent of the input content.
[0027] It should be noted that the target external service in this embodiment can refer to functions, data or resources provided by a third-party system, platform or component other than the large language model, for calling and using by the large language model. That is, the target external in this embodiment refers to functions, data or resources provided by a third-party system, platform or component capable of communicating with the large language model. As an example, the target external service can include a target external tool, which can be an external system or application that can be integrated with the large language model to provide additional functions or support. For example, the external tool can be a resource warehouse. As another example, the target external service can include a target data resource, for example, an external data resource can be a structured data resource or an unstructured data resource, and the like, which is not limited in this embodiment.
[0028] It should be noted that the large language model and the target external service in this embodiment can communicate with each other, for example, the large language model and the target external service can communicate based on a model context protocol (MCP).
[0029] It should be noted that the target operation can be any type of operation, for example, based on the input content, the target operation required to be performed by the target external service can be an operation of removing a preset number of goods, or an operation of adding a preset number of goods, and the like, which is not limited in this embodiment.
[0030] Step 103, using the large language model, calling the target external service to perform the target operation.
[0031] In some embodiments, the large language model can send an operation request to the target external service, wherein the operation request is used to instruct the target external service to perform the corresponding target operation. Correspondingly, after receiving the operation request, the target external service performs the corresponding target operation according to the operation request, obtains the operation result of the target operation, and returns the operation result to the large language model.
[0032] Step 104, using the large language model, generating reply content of the input content according to the operation result of the target operation performed by the target external service.
[0033] In this embodiment, after the large language model learns the operation result of the target operation performed by the target external service, the large language model can generate the reply content of the input content based on the operation result.
[0034] The information processing method based on the large language model provided in the embodiments of the present disclosure can determine, by the large language model, that the input content involves a target external service and a target operation required to be performed by the target external service, call the target external service to perform the target operation by the large language model, and generate reply content of the input content according to an operation result of the target external service performing the target operation by the large language model. In this way, the large language model can interact with the target external service involved in the input content, and accurately obtain the reply content of the input content based on the operation result of the target external service performing the corresponding target operation, thereby improving the accuracy of the reply content obtained by the large language model.
[0035] In some embodiments, in the case where the input content involves multiple target external services, whether the calling sequence of the multiple target external services is correct is very important for the large language model to accurately generate the reply of the input content. Therefore, in the embodiments, the multiple target external services can be sequentially called in combination with the calling sequence of the multiple external services, and the reply content of the input content can be generated by the large language model in combination with the operation results of the multiple target external services performing the corresponding target operations, respectively. In order to clearly describe the process, the process is exemplarily described below. Figure 2
[0036] Figure 2 is a schematic diagram according to the second embodiment of the present disclosure;
[0037] As shown in Figure 2 , the information processing method based on the large language model can include the following steps:
[0038] Step 201: obtaining input content.
[0039] The input content can be input by a user, for example, the input content can be input by a user in natural language.
[0040] Step 202: in the case where the input content involves multiple target external services, determining, by the large language model, a calling sequence of the multiple target external services and target operations required to be performed by the multiple target external services, respectively, by the large language model and the input content.
[0041] In the embodiments, the input content can be input to the large language model to determine, by the large language model, the calling sequence of the multiple target external services involved in the input content and the target operations required to be performed by the multiple target external services, respectively.
[0042] Step 203: sequentially calling, by the large language model, the multiple target external services to perform the corresponding target operations according to the calling sequence until the target operations fail or the multiple target external services successfully perform the corresponding target operations.
[0043] In this embodiment, in the process of using a large language model to call multiple target external services in sequence according to the calling order to perform target operations, for each called target external service, the large language model may send an operation request to the target external service, wherein the operation request is used to instruct the target external service to perform the corresponding target operation. Correspondingly, after receiving the operation request, the target external service performs the corresponding target operation according to the operation request, obtains the operation result of the target operation, and returns the operation result to the large language model, so that the large language model knows whether the corresponding target external service succeeds or fails to perform the corresponding target operation based on the operation result. If it is known that the corresponding target external service succeeds in performing the corresponding target operation, then an uncalled target external service is continued to be obtained according to the calling order. Correspondingly, the large language model continues to call the obtained target external service to perform the corresponding target operation until the target operation fails to be executed, or the above-mentioned multiple target external services all successfully perform the corresponding target operation.
[0044] Step 204 : Using the large language model, generate reply content to the input content according to the operation result of the target external service that successfully performs the target operation.
[0045] In some embodiments, a large language model may be used to generate reply content to the input content based on the operation results of the target external service that successfully performs the target operation and the operation results of the target external service that fails to perform the target operation.
[0046] In some embodiments, to further improve the effectiveness of the generated responses to the input content, a large language model is used to integrate the operation results of the target external services that successfully executed the target operation according to the call sequence to generate the response content to the input content. Thus, by integrating the operation results based on the order of the calls, the resulting response content to the input content is more logical and coherent, making the final response more reasonable and fluent.
[0047] The information processing method based on the large language model provided in the embodiments of the present disclosure, in the case that the input content involves multiple target external services determined by the large language model, determines the calling sequence of the multiple target external services and the target operation required to be executed by each of the multiple target external services by using the large language model and the input content; uses the large language model to sequentially call the multiple target external services to execute the corresponding target operation according to the calling sequence until the target operation fails or the multiple target external services all successfully execute the corresponding target operation; and uses the large language model to generate the reply content of the input content according to the operation result of the target external service that successfully executes the target operation. In this way, the large language model can sequentially call the external services to execute the corresponding target operation in combination with the calling sequence of the multiple external services involved in the input content, avoid logical confusion caused by the incorrect calling sequence of the multiple external services, and make the large language model able to accurately obtain the reply content of the input content based on the operation result of the target external service that successfully executes the target operation, thereby improving the accuracy of the reply content.
[0048] In some embodiments, in the case that it is determined that the input content does not involve the sequential connection words between the multiple target external services, the input content can be rewritten by using the large language model, so that the rewritten content involves the sequential connection words between the multiple target external services, and the calling sequence of the multiple target external services and the target operation required to be executed by each of the multiple target external services can be accurately obtained by analyzing the rewritten content by using the large language model. In order to clearly understand the process, the process is exemplarily described below. Figure 3
[0049] Figure 3 is a schematic diagram according to the third embodiment of the present disclosure.
[0050] As shown in Figure 3 , the method can further include the following steps:
[0051] Step 301, obtaining input content.
[0052] It should be noted that the specific description of step 301 can be referred to the related description in other embodiments, which will not be repeated here.
[0053] Step 302, in the case that the input content involves multiple target external services determined by using the large language model, rewriting the input content by using the large language model to obtain rewritten content, wherein the rewritten content includes sequential connection words between the multiple target external services.
[0054] In some embodiments, one possible implementation of using a large language model to determine that input content involves multiple target external services is to use the large language model to perform intent recognition on the input content to obtain multiple operational intents associated with the input content; and then select the external services in the external service list whose functional descriptions match the operational intents as target external services. This allows the large language model to accurately determine the multiple target external services associated with the input content based on the multiple operational intents of the input content.
[0055] It should be noted that, in this embodiment, different operation intentions may correspond to different target external services.
[0056] For example, if the external service is a warehouse, and the input content is: "I want to see the goods in warehouse 1 and give 100 to warehouse 2", the input content is analyzed by the large language model, and it can be determined that the input content includes two operation intentions, namely: operation intention Figure 1 and operating intention Figure 2 , among which, the operation intention Figure 1 Indicates the intention to operate warehouse 1. Figure 2 Indicates the intention to operate warehouse 2. Correspondingly, the function description and operation intention can be obtained from the warehouse list. Figure 1 Matching warehouses, and obtain function descriptions and operation intentions from the warehouse list Figure 2 Matching warehouse, assuming the functional description and operation intention Figure 1 The matching warehouse can be warehouse 1, function description and operation intention Figure 2 The matching warehouse may be warehouse 2.
[0057] In some embodiments, when the large language model determines that the input content involves multiple target external services, a first prompt word is generated based on the input content. The first prompt word is used to instruct the large language model to rewrite the input content, requiring that the rewritten content include sequential connectors between the multiple target external services. The first prompt word is then input into the large language model to generate the rewritten content. Thus, prompting the large language model with the first prompt word can guide the large language model to better understand the context and objectives of the generated content, thereby helping to improve the accuracy of the rewritten content generated by the large language model.
[0058] Among them, sequential conjunctions refer to words or phrases used to express a logical order of precedence. For example, sequential conjunctions may include but are not limited to first, second, next, then, subsequently, next, again, and finally. This embodiment does not specifically limit sequential conjunctions.
[0059] Step 303 : Using the large language model, determine the calling order of the multiple target external services and the target operations that each of the multiple target external services needs to perform according to the rewritten content.
[0060] In some embodiments, one possible implementation of step 303 is to use a large language model to determine the order in which the multiple target external services are called based on the sequential connectives between the multiple target external services in the rewritten content; and then, based on the large language model and the rewritten content, determine the target operations to be performed by each of the multiple target external services. Thus, by determining the order in which the multiple target external services are called and the target operations to be performed by each of the multiple target external services in two steps, more refined control is achieved, helping to improve execution stability.
[0061] For example, the rewritten content could be: First, call Warehouse 1 to reduce the quantity in Warehouse 1 by 100, then call Warehouse 2 to increase the quantity in Warehouse 2 by 100. It's understandable that whether Warehouse 1's quantity can be reduced by 100 depends on its current quantity. If Warehouse 1's current quantity exceeds 100, Warehouse 1 can reduce its quantity by 100. If Warehouse 1's current quantity does not exceed 100, Warehouse 1 cannot reduce its quantity by 100, and Warehouse 2 cannot increase its quantity by 100. In other words, only if Warehouse 1 successfully completes its target operation will Warehouse 2 be called to increase its quantity by 100.
[0062] Step 304 : Using the large language model, multiple target external services are called in sequence according to the calling order to perform corresponding target operations until the target operation fails to be executed, or multiple target external services all successfully perform the corresponding target operation.
[0063] Step 305 : Using the large language model, a reply content of the input content is generated according to the operation result of the target external service that successfully performs the target operation.
[0064] It should be noted that, for the specific description of step 304 and step 305, reference can be made to the relevant description in other embodiments, which will not be repeated here.
[0065] In this embodiment, when it is determined that the input content involves multiple target external services, the large language model can be used to rewrite the input content so that the rewritten content involves sequential connectives between multiple target external services. The large language model can accurately determine the calling order of multiple external target resources based on the sequential connectives between multiple target external services in the rewritten content, thereby helping to further improve the accuracy of the resulting reply content.
[0066] Figure 4 is a schematic diagram according to a fourth embodiment of the present disclosure.
[0067] like Figure 4 As shown, the method may further include the following steps:
[0068] Step 401: Obtain input content.
[0069] Step 402 : When the large language model is used to determine that the input content involves multiple target external services, the large language model and the input content are used to determine the calling order of the multiple target external services and the target operations that each of the multiple target external services needs to perform.
[0070] Step 403 : Using the large language model, multiple target external services are called in sequence according to the calling order to perform the corresponding target operation until the target operation fails to be executed, or multiple target external services all successfully perform the corresponding target operation.
[0071] Step 404 : Using the large language model, generate reply content to the input content according to the operation result of the target external service that successfully performs the target operation.
[0072] It should be noted that, for the specific description of steps 401 to 404, reference can be made to the relevant descriptions in other embodiments, which will not be repeated here.
[0073] Step 405, when the large language model is used to determine that the input content involves a target external service, the target operation that the target external service needs to perform is determined based on the large language model and the input content, and an operation request is sent to the target external service through the large language model, wherein the operation request is used to instruct the target external service to perform the target operation.
[0074] In some embodiments, the input content can be analyzed for intent using a large language model to obtain the operational intent of the input content, and a target external service whose functional description matches the operational intent can be obtained from the external service list. If it is determined that the number of the target external service is one, it can be determined that the input content involves one target external service.
[0075] In this embodiment, the input content may be analyzed using a large language model to determine the target operation that the target external service needs to perform based on the input content.
[0076] Step 406 : Generate transition text based on the input content and the functional description of the target external service using the large language model.
[0077] In some embodiments, a second prompt word is generated based on the input content and the functional description of the target external service. The second prompt word is used to instruct the large language model to generate transition text based on the input content and functional description. The second prompt word is then input into the large language model to generate transition text. Thus, the second prompt word can guide the large language model to generate transition text that meets the needs of the current scenario based on the input content and functional description, thereby improving the accuracy of the transition text generated by the large language model.
[0078] Step 407: Using the large language model, the operation result returned by the target external service is inserted after the transition text to obtain the reply content of the input content.
[0079] For example, if the input content is: "Let me see how much wheat is in warehouse 3," the large language model can determine that the target external service involved in this input content is warehouse 3, and the target operation required by warehouse 3 is to query the current amount of wheat. Correspondingly, the large language model can send a query request to warehouse 3, where the query request is for querying the current amount of wheat. Then, the input content and the functional description of warehouse 3 are input into the large language model to generate transition text through the large language model. Correspondingly, the large language model inserts the operation result returned by warehouse 3 after the transition text to obtain the response content of the input content and output the response content. Suppose the generated transition text is: "The grain has been stored. We have calculated: "," and the operation result returned by warehouse 2 is 2000 jin. Correspondingly, the response content of the input content obtained by the large language model based on the response content and the transition text can be: "The grain has been stored. We have calculated: 2000 jin."
[0080] In this embodiment, when the large language model is used to determine that the input content relates to a target external service, the target operation to be performed by the target external service is determined based on the large language model and the input content. After the large language model sends an operation request to the target external service, since it takes a certain amount of time for the external service to execute the corresponding target operation according to the operation request, in order to prevent the large language model from outputting text based on its own word segmentation units before receiving the operation result returned by the target external service, the large language model can be used to generate transition text based on the input content and the functional description of the target external service. The large language model is then used to insert the operation result returned by the target external service after the transition text to obtain the response content of the input content. This can improve the accuracy of the generated response content.
[0081] In order to implement the above embodiments, the present disclosure also provides an information processing device based on a large language model.
[0082] Figure 5 is a schematic diagram according to a fifth embodiment of the present disclosure.
[0083] like Figure 5 As shown, the information processing device 50 based on the large language model may include: an acquisition module 501, a first determination module 502, a calling module 503 and a generation module 504, wherein:
[0084] The acquisition module 501 is used to acquire input content.
[0085] The first determining module 502 is configured to determine, based on the large language model, whether the input content involves a target external service and a target operation that the target external service needs to perform.
[0086] The calling module 503 is used to use the large language model to call the target external service to perform the target operation.
[0087] The generation module 504 is configured to generate reply content to the input content based on the result of the target external service performing the target operation by using the large language model.
[0088] As a possible implementation of the embodiment of the present disclosure, when there are multiple target external services, the apparatus may further include:
[0089] The second determination module is used to determine the calling order of multiple target external services using the large language model and input content;
[0090] The calling module 503 is specifically configured to: use the large language model to call multiple target external services in sequence according to the calling order to perform corresponding target operations until the target operation fails to be performed, or multiple target external services all successfully perform the corresponding target operations;
[0091] The generation module 504 is specifically configured to use the large language model to generate reply content to the input content according to the operation result of the target external service that successfully performs the target operation.
[0092] As a possible implementation method of an embodiment of the present disclosure, the second determination module includes: a rewriting unit, which is used to rewrite the input content using a large language model to obtain rewritten content, wherein the rewritten content includes sequential conjunctions between multiple target external services; and a determination unit, which is used to determine the calling order of multiple target external services based on the rewritten content using the large language model.
[0093] As a possible implementation method of an embodiment of the present disclosure, the determination unit is specifically used to: generate a first prompt word based on the input content, wherein the first prompt word is used to instruct the large language model to generate rewritten content including sequential connectors between multiple target external services based on the input content; and input the first prompt word into the large language model to obtain the rewritten content.
[0094] As a possible implementation of the embodiment of the present disclosure, the determination unit is specifically configured to: utilize a large language model to determine a calling order of multiple target external services according to sequential connectives between the multiple target external services in the rewritten content.
[0095] As a possible implementation method of the embodiment of the present disclosure, the generation module 504 is specifically used to: use a large language model to integrate the operation results of the target external service that successfully executes the target operation according to the call order to obtain the reply content of the input content.
[0096] As a possible implementation method of the embodiment of the present disclosure, the first determination module 502 is specifically used to: use a large language model to identify the intent of the input content to obtain the operation intent of the input content; use the external service whose functional description matches the operation intent in the external service list as the target external service; and determine the target operation to be performed by the external service based on the large language model.
[0097] As a possible implementation of the embodiment of the present disclosure, when there is only one target external service, the apparatus further includes:
[0098] The transition text generation module is used to generate transition text based on the input content and the functional description of the target external service using a large language model;
[0099] The generation module 504 is specifically used to: use the large language model to insert the operation result returned by the target external service into the end of the transition text to obtain the reply content of the input content.
[0100] As a possible implementation of an embodiment of the present disclosure, a transition text generation module is specifically used to: generate a second prompt word based on the input content and the functional description of the target external service, wherein the second prompt word is used to instruct the large language model to generate transition text based on the input content and the functional description; and input the second prompt word into the large language model to obtain the transition text.
[0101] It should be noted that the aforementioned explanation of the embodiment of the information processing method based on a large language model is also applicable to the information processing device based on a large language model of this embodiment, and will not be repeated here.
[0102] The large language model-based information processing device of the disclosed embodiment uses the large language model to determine whether input content involves a target external service and a target operation required by the target external service. Using the large language model, the device then calls the target external service to perform the target operation. The large language model then generates a response to the input content based on the result of the target external service performing the target operation. This allows the large language model to interact with the target external service involved in the input content and, based on the result of the target external service performing the corresponding target operation, accurately obtain a response to the input content, thereby improving the accuracy of the response content obtained by the large language model.
[0103] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information are all carried out with the user's consent, comply with relevant laws and regulations, and do not violate public order and good morals.
[0104] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.
[0105] Figure 6 is a block diagram of an electronic device for implementing the information processing method based on a large language model of an embodiment of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.
[0106] like Figure 6 As shown, the electronic device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. Various programs and data required for the operation of the electronic device 600 can also be stored in the RAM 603. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.
[0107] Multiple components in the electronic device 600 are connected to the I / O interface 605, including an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a magnetic disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the electronic device 600 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.
[0108] The computing unit 601 can be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the information processing method based on the large language model. For example, in some embodiments, the information processing method based on the large language model can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the information processing method based on the large language model described above can be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to execute the information processing method based on the large language model in any other appropriate manner (for example, by means of firmware).
[0109] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0110] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.
[0111] In the context of the present disclosure, a machine-readable medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0112] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).
[0113] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.
[0114] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact through a communication network. The client-server relationship arises through computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, a server in a distributed system, or a server integrated with a blockchain.
[0115] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not a limitation herein.
[0116] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.
Claims
1. An information processing method based on a large language model, comprising: Get input content; determining, based on the large language model, that the input content relates to a target external service and a target operation to be performed by the target external service; Using the large language model, calling the target external service to perform the target operation; The large language model is used to generate reply content for the input content according to an operation result of the target external service executing the target operation.
2. The method according to claim 1, wherein In the case where there are multiple target external services, the method further includes: Determining a calling order of the plurality of target external services by using the large language model and the input content; The step of utilizing the large language model to call the target external service to perform the target operation includes: Using the large language model, calling the plurality of target external services in sequence according to the calling order to perform the corresponding target operation until the target operation fails to be performed, or the plurality of target external services all successfully perform the corresponding target operation; The step of utilizing the large language model to generate reply content to the input content according to the operation result of the target external service executing the target operation includes: The large language model is used to generate reply content for the input content according to an operation result of the target external service that successfully performs the target operation.
3. The method according to claim 2, wherein: The determining the calling order of the plurality of target external services by using the large language model and the input content includes: rewriting the input content using the large language model to obtain rewritten content, wherein the rewritten content includes sequential connectives between the plurality of target external services; The large language model is used to determine a calling order of the plurality of target external services according to the rewritten content.
4. The method according to claim 3, wherein: The rewriting of the input content by using the large language model to obtain rewritten content includes: generating a first prompt word according to the input content, wherein the first prompt word is used to instruct the large language model to generate rewritten content including sequential connectives between the plurality of target external services according to the input content; The first prompt word is input into the large language model to obtain the rewritten content.
5. The method according to claim 3, wherein The operation of using the large language model to determine the calling sequence of the plurality of target external services according to the rewritten content includes: The large language model is used to determine a calling order of the plurality of target external services according to sequential connectives between the plurality of target external services in the rewritten content.
6. The method according to claim 2, wherein: The step of utilizing the large language model to generate reply content to the input content according to an operation result of the target external service that successfully performs the target operation includes: The large language model is used to integrate the operation results of the target external service that successfully executes the target operation according to the calling sequence to obtain the reply content of the input content.
7. The method according to any one of claims 1 to 6, wherein The determining, based on the large language model, that the input content involves a target external service and a target operation to be performed by the target external service includes: Using the large language model to perform intent recognition on the input content to obtain the operation intent of the input content; The external service in the external service list whose function description matches the operation intention is used as the target external service; A target operation to be performed by the external service is determined based on the large language model.
8. The method according to claim 1, wherein In the case where there is one target external service, after using the large language model to call the target external service to perform the target operation, the method further includes: generating transition text based on the input content and the functional description of the target external service using the large language model; Generating reply content to the input content based on an operation result of the target external service executing the target operation using the large language model, including: The operation result returned by the target external service is inserted after the transition text using the large language model to obtain reply content of the input content.
9. The method according to claim 8, wherein The step of generating transition text based on the input content and the functional description of the target external service by using the large language model includes: generating a second prompt word according to the input content and the functional description of the target external service, wherein the second prompt word is used to instruct the large language model to generate a transition text based on the input content and the functional description; The second prompt word is input into the large language model to obtain the transition text.
10. An information processing device based on a large language model, comprising: Acquisition module, used to obtain input content; A first determining module is configured to determine, based on a large language model, whether the input content involves a target external service and a target operation to be performed by the target external service; A calling module, configured to use the large language model to call the target external service to perform the target operation; A generation module is used to generate reply content of the input content according to the operation result of the target external service performing the target operation by using the large language model.
11. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 9.
12. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 9.
13. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 9.