Form filling method, system and device, storage medium and program product

Abstract

The embodiment of the invention provides a form filling method, system and device, a storage medium and a program product. The method comprises the following steps that a target interface is displayed, the target interface comprises a to-be-filled form, and the to-be-filled form comprises a first to-be-filled item and a first control; in response to a trigger operation for the first control, displaying a first question proposed to a virtual assistant on the target interface; and displaying a first reply of the virtual assistant for the first question on the target interface, wherein the first reply comprises first recommendation information used for filling the first to-be-filled item. According to the technical scheme provided by the embodiment of the invention, the form filling efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to a form filling method, system, device, storage medium and program product. Background Art

[0002] With the continuous development of Internet technology, the application of the Internet has become very common. More and more users can carry out various activities through the Internet, and in this process, they will encounter situations where they need to fill in online forms.

[0003] For example: in the process of online trade (especially online international trade), when a buyer files a chargeback application for a certain order with a bank, the bank will require the merchant to provide evidence to defend the chargeback application initiated by the buyer within a specified time. The bank will review the defense materials provided by the merchant and make a final decision. In order to assist merchants in providing evidence and defense, the e-commerce platform will provide merchants with an online form for collecting defense materials to be filled out by the merchants, and then submit a chargeback defense application to the bank based on the defense materials filled out by the merchants.

[0004] However, currently the efficiency of filling out online forms is low. Summary of the invention

[0005] In view of the above problems, the present application is proposed to provide a form filling method, system, device, storage medium and program product that solve the above problems or at least partially solve the above problems.

[0006] In a first aspect of the present application, a form filling method is provided, applicable to a terminal, comprising:

[0007] Displaying a target interface, wherein the target interface includes a form to be filled in, and the form to be filled in includes a first item to be filled in and a first control;

[0008] In response to a trigger operation on the first control, displaying a first question posed to the virtual assistant on the target interface;

[0009] The first response of the virtual assistant to the first question is displayed on the target interface, wherein the first response includes first recommended information for filling in the first item to be filled in.

[0010] A second aspect of the present application provides a form filling method, applicable to a server, comprising:

[0011] Receiving a first target request sent by a terminal, the first target request including first indication information and second prompt information, the first indication information is used to indicate a first to-be-filled item of a to-be-filled form, the second indication information is used to indicate a target object, and the to-be-filled form is used to collect information related to the target object;

[0012] Determining, according to the target object, first recommended information for filling in the first item to be filled in by using a large language model;

[0013] The first recommendation information is returned to the terminal.

[0014] In a third aspect of the present application, a form filling system is provided, comprising: a terminal and a server; wherein:

[0015] The terminal is used to: display a target interface, the target interface includes a form to be filled in, the form to be filled in is used to collect information related to a target object, and the form to be filled in includes a first item to be filled in and a first control; in response to a trigger operation on the first control, send a first target request to the server and display a first question posed to the virtual assistant on the target interface, the first target request includes first indication information and second indication information, the first indication information is used to indicate the first item to be filled in, the second indication information is used to indicate the target object, and the first target request is used to request the server to provide the first recommendation information based on the target object using the large language model;

[0016] The server is used to: receive a first target request sent by the terminal; determine the first recommendation information according to the target object using a large language model; and return the first recommendation information to the terminal.

[0017] The terminal is further used to: receive the first recommendation information returned by the server; and display the first reply of the virtual assistant to the first question on the target interface, wherein the first reply includes the first recommendation information.

[0018] In a fourth aspect of the present application, an electronic device is provided. The electronic device comprises: a memory and a processor, wherein:

[0019] The memory is used to store programs;

[0020] The processor is coupled to the memory, and is used to execute the program stored in the memory to implement any of the methods described above.

[0021] According to a fifth aspect of the present application, a computer-readable storage medium storing a computer program is provided, wherein the computer program can implement any of the above-described methods when executed by a computer.

[0022] In a sixth aspect of the present application, a computer program product is provided, comprising a computer program, which implements any of the methods described above when executed by a processor.

[0023] The embodiment of the present application provides an accompanying data filling solution, that is, when the user is filling out a form, a virtual assistant (or virtual robot) is used to provide relevant filling recommendations for the items to be filled in the form, thereby improving the filling efficiency and reducing the time spent on filling out. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0025] Figure 1 A schematic diagram of model iteration provided in an embodiment of the present application;

[0026] Figure 2 A flowchart of a form filling method provided in an embodiment of the present application;

[0027] Figure 3 A schematic diagram of a chargeback details interface provided in an embodiment of the present application;

[0028] Figure 4 A schematic diagram of a form filling interface provided in an embodiment of the present application;

[0029] Figure 5 A flowchart of a form filling method provided in another embodiment of the present application;

[0030] Figure 6 A structural diagram of a form filling system provided in yet another embodiment of the present application;

[0031] Figure 7 A schematic diagram of the structure of a form filling solution provided in an embodiment of the present application;

[0032] Figure 8 A structural block diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0033] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below according to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0034] In addition, some of the processes described in the specification, claims and the above-mentioned figures of the present application include multiple operations that appear in a specific order, and these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0035] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0036] First, the vocabulary involved in the embodiments of the present application is explained. It can be understood that this explanation is for a clearer understanding of the embodiments of the present application, and does not necessarily constitute a limitation on the embodiments of the present application.

[0037] Chargeback: Also known as return or withdrawal, it refers to the credit card holder (buyer) applying to the bank (credit card issuing organization) to refuse to pay a certain order on the bill within a certain period after payment. Chargeback is actually a right granted by the bank to the credit card holder. No matter what e-commerce platform the merchant (seller) uses, such risks cannot be completely avoided. When a buyer pays for an online transaction on an e-commerce platform by credit card, the buyer can file a chargeback with the bank according to the rules and time limit set by the bank. The dispute over chargeback will be ultimately decided by the bank.

[0038] Chargeback defense: The bank requires the merchant to provide evidence to defend the chargeback application initiated by the buyer within a specified time. The bank will review the defense materials provided by the seller and make a final decision. Among them, the timeliness and professionalism of the preparation of the defense materials determine whether the chargeback defense can win.

[0039] Comprehensive chargeback service (or chargeback guarantee service): In order to better protect the rights and interests of merchants and reduce the losses caused by chargeback risks to merchants, e-commerce platforms provide comprehensive chargeback services to merchants who have opened credit guarantee services. If the conditions stipulated in the comprehensive chargeback service are met and the merchant's comprehensive chargeback service quota is sufficient, regardless of the final defense result, the e-commerce platform will bear the chargeback refund amount loss for the merchant within the merchant's comprehensive chargeback service quota.

[0040] A large language model (LLM) refers to a deep learning model trained with a large amount of text data that can generate and understand natural language text. Such models are usually composed of billions or even hundreds of billions of parameters and are trained through deep neural networks to handle complex natural language tasks such as text classification, question and answer, and dialogue.

[0041] Prompts are a way to interact with the LLM model. They guide the model to produce the desired output through language instructions. Prompts can be questions, instructions, or any form of text input, which are designed to stimulate the model to generate a specific type of response or complete a specific task.

[0042] The prompt word template is a pre-designed prompt word framework with a certain structure and format. It includes fixed parts and placeholders. The placeholders in the template are positions to be filled. By filling in the placeholders, different prompt words can be generated. The prompt word template helps to improve the efficiency and standardization of prompt word generation, making the interaction with LLM more convenient and effective.

[0043] In order to improve the efficiency of filling out online forms (hereinafter referred to as forms), an embodiment of the present application provides an accompanying data filling solution, that is, in the process of users filling out forms, a virtual assistant (or virtual robot) is used to provide relevant filling recommendations for the items to be filled in the form, thereby improving the filling efficiency and reducing the time spent on filling out.

[0044] The virtual assistant mentioned in the embodiment of the present application is a software based on artificial intelligence (AI) technology, which can recognize and understand the user's language input, search and process information, and output appropriate responses.

[0045] In an optional implementation, the virtual assistant may be implemented based on the LLM model.

[0046] Generally, the LLM model involves a pre-training phase and a fine-tuning phase. The pre-training phase is the initial stage of LLM model learning. During the pre-training phase, the model is exposed to a large amount of unlabeled text data, such as books, articles, and websites. The purpose is to capture the underlying patterns, structures, and semantic knowledge present in the text corpus, and to obtain a model with preliminary general knowledge or capabilities through training using a large amount of unlabeled or weakly labeled data. The purpose of pre-training is to allow the model to capture widely useful features by learning a large amount of general data before performing a specific task, thereby improving the model's performance and generalization ability on specific tasks.

[0047] Although the pre-trained model has learned a wealth of general features and prior knowledge from large-scale datasets, these features and knowledge may not be fully applicable to specific target tasks. Fine-tuning further trains the pre-trained model on a small amount of labeled data related to a specific task, allowing the model to learn specific features and rules related to the target task, thereby better adapting to the specific task.

[0048] In addition to improving the performance of LLM models through pre-training and fine-tuning, the performance of LLM models can also be improved through technologies such as prompt engineering and retrieval-augmented generation (RAG). Among them, prompt engineering is a method specifically for optimizing language models, that is, a method for optimizing prompt templates. Its goal is to guide these models to generate more accurate and targeted output texts by designing and adjusting prompt templates. Retrieval-augmented generation is a technology that combines retrieval and generation. It generates answers or content by referencing information from an external knowledge base (also known as a plug-in knowledge base). It has strong interpretability and customization capabilities and is suitable for multiple natural language processing tasks such as question-answering systems, document generation, and intelligent assistants. The advantages of RAG-based models are strong versatility, instant knowledge updates, and more efficient and accurate information services through end-to-end evaluation methods.

[0049] Optionally, the LLM model can be obtained based on an existing pre-trained model, for example, based on an existing Generative Pre-trained Transformer (GPT) model. In this way, the pre-training stage can be skipped, and the ability of the pre-trained model on a specific task can be improved directly through one or more of the following methods: prompt word engineering, retrieval enhancement generation, and fine-tuning.

[0050] Usually, a form involves multiple items to be filled in, and each item requires different content. Optionally, in order to improve the model's ability to provide filling recommendations, different prompt word templates can be configured for different items to be filled in the form. The prompt word template corresponding to each item to be filled in can be obtained based on prompt word engineering optimization.

[0051] Taking the chargeback defense scenario as an example, training corpus can be generated based on historical defense cases of historical orders. The training corpus may include: the defense materials submitted in the historical defense cases and the order information of historical orders. The defense materials are the defense materials collected based on the form in history, which include the filled-in information of each item to be filled in the form, and the order information of historical orders, including but not limited to: order content, buyer information and seller information, logistics information and vouchers, payment information and vouchers, the reasons for the buyer's chargeback against the order, and the communication records between the buyer and the seller regarding the order. Among them, the logistics vouchers may include shipping vouchers and delivery vouchers. Among them, the training corpus generated based on the historical defense cases with successful defense can be used as positive corpus, and the training corpus generated based on the historical defense cases with failed defense can be used as negative corpus.

[0052] According to the positive corpus and the negative corpus, the prompt word template corresponding to each item to be filled in can be tuned / debugged to obtain a prompt word template that meets the preset requirements. In this way, when providing filling recommendations for users later, the filling recommendations provided by LLM for the items to be filled in can be obtained based on the prompt word target. The prompt word template tuning process can be implemented based on the existing prompt word project, and the embodiment of the present application does not specifically limit this.

[0053] In actual applications, after the user (i.e., the merchant) fills out the form and submits it to the e-commerce platform, the e-commerce platform can use the LLM model to evaluate whether the defense information collected in the form can be successfully defended. In other words, before the e-commerce platform initiates a chargeback defense application to the bank, the electronic commodity platform first estimates whether the defense information submitted by the user can be successfully defended. In one example, the e-commerce platform can determine whether to provide the user with a chargeback protection service based on the estimated results. In another example, the e-commerce platform can determine whether it is necessary to remind the user to modify the form based on the estimated results to increase the success rate of the defense.

[0054] In order to improve the performance of the model in the task of evaluating whether a defense is successful, the prompt words, external knowledge base and / or fine-tuning model can be optimized based on positive and negative corpora.

[0055] In one embodiment, the prompt word template can be tuned / debugged based on the positive corpus and the negative corpus to obtain a prompt word template that meets the preset requirements. In this way, when evaluating whether the defense materials collected by the form can be successful, the evaluation results of the LLM model can be obtained based on the prompt word template.

[0056] In another embodiment, an external knowledge base (also known as a plug-in knowledge base) is mounted on the model, and the knowledge base includes positive corpus and negative corpus, wherein the defense result recorded corresponding to the positive corpus is a successful defense, and the defense result recorded corresponding to the negative corpus is a failed defense. In this way, when evaluating whether the defense data collected by the form can be successful, the defense data collected by the form can be matched with the positive corpus and negative corpus in the knowledge base. If they match, the defense result corresponding to the matched corpus can be used as the evaluation result; if they do not match, the large model generates the evaluation result based on the parameters of the model.

[0057] In another embodiment, the parameters of the model can be fine-tuned according to the positive corpus and the negative corpus to obtain a model that meets the preset performance requirements. In the process of fine-tuning the model, the training corpus can be divided into a training set and a test set, wherein the training set is used to fine-tune the parameters of the model, and the test set is used to test or evaluate the prediction accuracy of the model. When the prediction accuracy is greater than or equal to the preset threshold, it means that the performance of the model has met the preset requirements and the fine-tuning can be stopped. The existing model fine-tuning method can be used to fine-tune the model, and the embodiment of the present application does not specifically limit this.

[0058] In practical applications, one or more methods of optimizing prompt word templates, plug-in knowledge bases, and fine-tuning models can be selected to improve the performance of the model, and the embodiments of the present application do not specifically limit this. For example, a large model plug-in knowledge base can be first used; then the prompt word template can be optimized to obtain an optimized prompt word template; finally, the model can be fine-tuned based on the optimized prompt word template. It should be noted that when the performance of the model in the task of evaluating whether the defense can be successfully achieved by optimizing the prompt word template meets the preset performance requirements, the model does not need to be fine-tuned; when the performance of the model in the task of evaluating whether the defense can be successfully achieved by optimizing the prompt word template cannot meet the preset performance requirements, the model can continue to be fine-tuned.

[0059] In addition, after the model is put into use, it can be fine-tuned based on the newly generated defense cases, so as to iterate a better model. For example, new training corpus can be generated based on the newly generated defense cases, and the new training corpus can be used as the test set of the model. By comparing the evaluation results of the model on the test set with the actual defense results of the test set, the model accuracy of the model on the test set can be evaluated; if the accuracy is lower than the preset threshold, the model can be fine-tuned based on the new training corpus and the training corpus used to fine-tune the model in the past, so as to iterate a better model.

[0060] like Figure 1As shown, the model is continuously iterated using the newly generated training corpus, where the performance of model V2 is higher than that of model V1, and the performance of model V3 is higher than that of model V2.

[0061] Of course, the prompt word template can be tuned and the knowledge of the plug-in knowledge base can be updated based on the newly generated training corpus, for example, the newly generated training corpus can be added to the plug-in knowledge base.

[0062] Usually, the number of items to be filled in a form is large and time-consuming. In order to improve the efficiency of form filling, this application provides a form filling method, which is applicable to terminals such as Figure 2 As shown, the method comprises the following steps:

[0063] 201. Display the target interface.

[0064] Wherein, the target interface includes a form to be filled in.

[0065] Exemplarily, the form to be filled in is used to collect information related to the target object. In different application scenarios, the target object is different. For example, if the form to be filled in is an annual report of an enterprise, then the target object is the target enterprise. For example, if the form to be filled in is a defense information form, the defense information form is used to collect defense information for a chargeback application. The chargeback application is filed against a target order, then the target object refers to the target order.

[0066] In different application scenarios, the information required to be collected in the form to be filled in is different. Therefore, the items to be filled in the form to be filled in can be designed according to different application scenarios.

[0067] Take the defense against chargeback scenario as an example. Figure 4 The form filling interface (also known as the target interface) 40 shown includes a form 41 to be filled in and a conversation box 42 with a chargeback defense assistant (also known as a virtual assistant).

[0068] In an optional implementation, in response to a user (i.e., a merchant) clicking on a target order in an order list interface, an order details interface of the target order is displayed, and the order details interface includes a refund management control; in response to a user triggering an operation on the refund management control, a refund management interface is entered, and the refund management interface includes a chargeback control. In response to a user triggering an operation on the chargeback control, a chargeback details interface 30 (e.g., Figure 3 As shown in Figure 3 As shown, the chargeback details interface 30 may include a fill-in defense information control 31. In response to the user's triggering operation on the fill-in defense information control 31, the display Figure 4 The form filling interface 40 shown. Optionally, Figure 3As shown, the chargeback details interface 30 may also include: order information 32 of the target order, chargeback details information 33. The order information may include order number, order total price, prepayment information and balance information. The chargeback details information may include: chargeback type, chargeback reason, chargeback amount and chargeback time. Figure 3 As shown, the chargeback details interface 30 may also include: an operation record 34 related to the chargeback.

[0069] For example, Figure 4 As shown, the form 41 to be filled in is located in the left area of ​​the form filling interface 40 , and the conversation box 42 is located in the right area of ​​the form filling interface 40 .

[0070] It should be noted that the layout of the form to be filled in and the dialog box on the target interface can be set according to actual needs, and the embodiment of the present application does not specifically limit this. For example: the form to be filled in is located in the upper area of ​​the form filling interface, and the dialog box is located in the lower area of ​​the form filling interface; for another example: the form to be filled in is located in the right area of ​​the form filling interface, and the dialog box 42 is located in the left area of ​​the form filling interface 40.

[0071] like Figure 4 As shown, the form 41 to be filled in may include multiple items to be filled in. Taking the chargeback defense scenario as an example, Figure 4 As shown, the form 41 to be filled in may include: an item to be filled in for entering reasons for defense, an item to be filled in for entering the return and exchange address, an item to be filled in for entering the logistics order number and tracking website, an item to be filled in for entering the delivery time, an item to be filled in for entering communication proof materials such as emails and instant chats, an item to be filled in for filling in proof of shipment, an item to be filled in for filling in proof of delivery, and an item to be filled in for filling in the closing statement.

[0072] In actual applications, some of the items to be filled in are used to fill in text information, and some of the items to be filled in are used to add files.

[0073] When the item to be filled in is used to fill in text information, the item to be filled in may include an input box (such as Figure 4 As an option, when the item to be filled in has the attribute of filling in multiple input information, the item to be filled in may include multiple input boxes, or the item to be filled in may include an add control for adding an additional input box, for example: Figure 4 The items to be filled in for filling in the logistics order number and tracking website have the attribute of filling in multiple filled-in information, which corresponds to Figure 4 The "add order number" control 416 is shown. In response to the trigger operation for adding the control, an input box can be added to the target interface for the item to be filled in.

[0074] like Figure 4As shown, when the item to be filled in can be filled in with the assistance of a virtual assistant, a corresponding assistance control (such as Figure 4 412 shown). When there are multiple items to be filled in that can be filled in with the assistance of a virtual assistant, a corresponding assistance control can be set for each item to be filled in the form to be filled in, and a corresponding relationship between the assistance control and the item to be filled in can be established, and different items to be filled in correspond to different assistance controls. Exemplarily, the assistance control corresponding to each item to be filled in is displayed in association with the item to be filled in, wherein the associated display can be close to the display position or the two are connected by an interface element for association. Figure 4 As shown, a to-be-filled item 411 for filling in the defense reason is displayed close to its corresponding assisting control 412 .

[0075] Optionally, the item to be filled in has an attribute for receiving manual filling by the user. That is, the terminal can receive information filled in by the filling operation in response to the filling operation on the item to be filled in, and use the information as the filling information of the item to be filled in. In this way, when the virtual assistant cannot provide filling recommendations or provides incorrect filling recommendations, the user can manually fill in the relevant content to ensure the accuracy of the filling.

[0076] The form to be filled in includes a first item to be filled in and a first control. The first item to be filled in can be any item to be filled in the form, and the first control refers to an auxiliary control corresponding to the first item to be filled in.

[0077] 202. In response to a trigger operation on a first control, display a first question posed to the virtual assistant on the target interface.

[0078] The first question is used to inquire about filling recommendations for the first item to be filled in.

[0079] Optionally, in response to the triggering operation on the first control, the item to be filled in (i.e., the first item to be filled in) corresponding to the first control is determined according to the correspondence between the preset control and the item to be filled in, and the first question is determined according to the first item to be filled in. Exemplarily, a corresponding question can be configured for each item to be filled in in advance, and a correspondence between the two can be established. Subsequently, according to the correspondence, the question corresponding to the first item to be filled in can be determined as the first question. After the first question is determined, the first question posed to the virtual assistant is displayed on the target interface.

[0080] Exemplarily, the trigger operation may include: a click operation.

[0081] For example, Figure 4As shown, in response to the triggering operation of the assistance control 412 corresponding to the to-be-filled item 412 for filling in the defense reason, the question 421 posed to the virtual assistant is displayed in the conversation box 42 of the target interface: "Recommend me a suitable defense reason for this defense."

[0082] In an embodiment of the present application, the user only needs to click the control corresponding to the item to be filled in to input questions to the virtual assistant for filling recommendations for the item to be filled in. The user does not need to manually edit the questions, which reduces the user's usage cost and improves filling efficiency.

[0083] 203. Displaying a first response of the virtual assistant to the first question on the target interface.

[0084] The first reply may include first recommendation information for filling in the first item to be filled in.

[0085] For example, Figure 4 As shown, the virtual assistant's answer 422 to the question 421 may be displayed in the conversation box 42 of the target interface.

[0086] Optionally, the first reply may also include the basis for the recommendation or the reason for the recommendation, such as Figure 4 As shown, the reply 422 includes “Recommendation reason introduction: ***”. In this way, the user can judge whether the recommendation of the virtual assistant is accurate based on the recommendation reason.

[0087] Subsequently, the user can complete the filling of the first item to be filled in based on the first reply. For example, the user can copy the first recommendation information and paste it into the first item to be filled in, thereby completing the filling task of the first item to be filled in.

[0088] In the technical solution provided in the embodiment of the present application, when a user is filling out a form, the user can use the virtual assistant to provide relevant filling recommendations for the items to be filled in the form by clicking the corresponding controls, which helps to improve filling efficiency and save time.

[0089] Applying the technical solution provided in the embodiment of the present application in the scenario of defense against payment rejection can also help improve the professionalism of the filled-in content, thereby helping to improve the success rate of the defense.

[0090] In an optional implementation, the virtual assistant may include an LLM model.

[0091] It should be noted that, in some embodiments, the LLM model of the virtual assistant runs on the server side, so the terminal can obtain the filling recommendation of the LLM model for the first item to be filled in by interacting with the server side, and then generate the first answer of the virtual assistant to the first question based on the filling recommendation. In other embodiments, when the hardware resources of the terminal meet the operating requirements of the LLM model, the LLM model of the virtual assistant can be deployed on the terminal, so that the virtual assistant on the terminal uses the locally deployed LLM model to obtain the filling recommendation for the first item to be filled in, and generates the first answer of the virtual assistant to the first question based on the filling recommendation.

[0092] In one embodiment, the first reply may also include a recommendation basis or a recommendation reason, wherein the recommendation basis and the recommendation reason may be provided by the LLM model. By displaying the recommendation basis or the recommendation reason, the user can further judge whether the model recommendation is accurate. If it is accurate, the user adopts the model recommendation content; if it is inaccurate, the user manually inputs it.

[0093] When the first reply includes the recommendation basis corresponding to the first recommendation information, a fourth control is displayed on the target interface; in response to a trigger operation on the fourth control, a detail interface is displayed, wherein the detail interface is used to display detailed information about the recommendation basis.

[0094] like Figure 4 As shown, the target interface 40 may display a fourth control 423; in response to the triggering operation on the fourth control 423, a detail interface (not shown) is displayed, and the detail interface is used to display detailed information about the recommendation basis. In this way, the user can conveniently view the detailed information of the recommendation basis. For example, Figure 4 As shown, the recommendation basis for "logistics order number and tracking website" includes logistics voucher. In response to the user's click operation on the control "click to view" 423, the details interface is displayed, and the details interface can display information such as logistics voucher pictures.

[0095] In an optional implementation, the above method may further include:

[0096] 204. Display a second control on the target interface.

[0097] Exemplarily, the first recommendation information and the second control are displayed in a preset area of ​​the target interface.

[0098] Wherein, in the target interface, the preset area is associated with the area where the first item to be filled in is located. Herein, the association of two areas means that the areas are close to each other or the areas are associated through interface elements.

[0099] For example, Figure 4As shown, in the target interface, the defense reason 413 (ie, the first recommended information) recommended by the virtual assistant and the confirmation filling control (ie, the second control) 414 are displayed near the to-be-filled item 411 (ie, the first to-be-filled item) for filling in the defense reason.

[0100] In one embodiment, when the first item to be filled in is used to add a file, the first recommendation information includes a first recommended file, where the file can be a document or a picture. When the first recommended file is a document file, the file name of the document file is displayed on the target interface; when the first recommended file is a picture, a thumbnail of the picture (such as a thumbnail) can be displayed. Figure 4 Thumbnail 432 shown) and / or picture name (such as Figure 4 The first recommended file may be one or more files. When there are multiple first recommended files, the second control (such as Figure 4 The "application" control 434 shown is also multiple, and the multiple second controls correspond one-to-one to the multiple first recommended files. Different first recommended files correspond to different second controls, and the corresponding first recommended files and second controls can be associated and displayed on the target interface.

[0101] Optionally, when the number of first recommended files is greater than a preset threshold (for example, 2), part of the first recommended files and an expansion control 435 may be displayed on the target interface, and in response to a trigger operation on the expansion control 435, the first recommended files that have not yet been displayed in the form area may be displayed.

[0102] 205. In response to a triggering operation on the second control, fill in the first recommended information into the first item to be filled in.

[0103] Exemplarily, the trigger operation may be a click operation.

[0104] When the first item to be filled in includes a first input box (that is, the first item to be filled in is used to fill in text information), in response to a trigger operation on the second control, the first recommendation information is input into the first input box.

[0105] Optionally, the first recommendation information is displayed in the first input box.

[0106] Optionally, after the first recommendation information is filled in the first item to be filled in, the first recommendation information and the second control may disappear from the target interface, that is, the first recommendation information and the second control are not displayed on the target interface.

[0107] like Figure 4 As shown, the items to be filled in for filling in the logistics order number and the tracking website include an input box 417 and an add control 416 for adding an additional input box. When the virtual assistant recommends multiple pieces of recommended information for the items to be filled in (for example: Figure 4When the logistics information 1 and logistics information 2 are displayed, the recommended information 419 (for example: Figure 4 The recommended information 419 is displayed as logistics information 1, and a "confirm fill-in" control 418 corresponding to the recommended information 419 is displayed. In response to the trigger operation on the "confirm fill-in" control 418, the recommended information 419 is filled into the input box 417. In response to the trigger operation on the add control 416, an additional input box (not shown) is added for the item to be filled in on the target interface, and other recommended information among the multiple recommended information is displayed on the target interface (other recommended information refers to recommended information that has not been filled in, for example: Figure 4 The logistics information 2) shown and its corresponding "confirm fill-in" control (not shown) respond to the trigger operation on the control, and fill in the other recommended information into the additional input box.

[0108] Exemplarily, on the target interface, other recommended information and its corresponding "confirm fill in" control are displayed in the plurality of recommended information in a manner of replacing the recommended information 419 and its corresponding "confirm fill in" control 418. After the replacement, the target interface no longer displays the recommended information 419 and its corresponding "confirm fill in" control 418, but displays other recommended information and its corresponding "confirm fill in" control.

[0109] When the first item to be filled in is used to add a file, the first recommendation information includes a first recommended file, and in response to a trigger operation on the second control, the first recommended file is used as an added file for the first item to be filled in.

[0110] Optionally, on the target interface, the added files of the first item to be filled in are displayed, for example: Figure 4 As shown, a thumbnail 431 of the added picture of the first item to be filled in is displayed, or the file name of the added file is displayed.

[0111] In an embodiment of the present application, after the virtual assistant gives a filling recommendation, the user can fill in the recommended information into the item to be filled in by triggering an operation through a control, thereby further improving the filling efficiency and the user filling experience.

[0112] In one embodiment, the virtual assistant includes a large language model running on the server. The above method may also include the following steps:

[0113] 206. In response to the triggering operation on the first control, send a first target request to the server.

[0114] Among them, the first target request includes first indication information and second indication information, the first indication information is used to indicate the first item to be filled in, and the second indication information is used to indicate the target object. The first target request is used to request the server to provide the first recommendation information based on the target object using the large language model.

[0115] Exemplarily, the first indication information may include the name of the first item to be filled in. Exemplarily, the name of the item for filling in the defense reason is the defense reason.

[0116] Exemplarily, the second indication information may include an object identifier of the target object. Taking the target order as an example, the object identifier may be an order number.

[0117] Exemplarily, the first target request is generated by the terminal in response to a trigger operation on the first control.

[0118] In an optional implementation manner, after receiving the first target request, the server performs the following steps:

[0119] S1. Determine, according to the target object, first recommended information for filling in the first item to be filled in by using a large language model.

[0120] S2. Return the first recommendation information to the terminal.

[0121] In the above S1, object information of the target object may be obtained, a first target prompt word may be generated according to the object information of the target object, and the first target prompt word may be input into the LLM model to obtain first recommendation information. The first target prompt word is used to prompt the large language model to provide first recommendation information for filling in the first to-be-filled item according to the object information.

[0122] Taking the chargeback defense scenario as an example, the target object is the target order, then the object information is the order information, which includes but is not limited to: order content, buyer information and seller information, logistics information, logistics voucher, payment information, payment voucher, buyer's chargeback reason for the order, and communication records between the buyer and seller regarding the order. Among them, the order content may include but is not limited to: product type, product name, product quantity, product unit price, and product total price.

[0123] It should be noted that for different items to be filled in, the object information required to be obtained is different.

[0124] Take the chargeback defense scenario as an example:

[0125] For the items to be filled in for filling in the reasons for defense, you can obtain order information such as the reason for chargeback, logistics information, payment information, and product type.

[0126] For the unfilled items for filling in the logistics order number and tracking URL, the logistics information can be obtained.

[0127] For the items to be filled in for entering the delivery time, logistics information can be obtained.

[0128] For shipping pictures, the communication records between buyers and sellers can be obtained, and the pictures that have been uploaded can be obtained from the records. Among them, shipping pictures refer to logistics pictures related to shipping.

[0129] A method for constructing a first target prompt word is described in detail below. The method includes the following steps:

[0130] S11. Obtain a first preset prompt word template adapted to the first item to be filled in.

[0131] According to the correspondence between the preset items to be filled in and the preset prompt word templates, a first preset prompt word template adapted to the first item to be filled in can be obtained.

[0132] Different items to be filled in correspond to different preset prompt word templates. The process of determining / optimizing the preset prompt word template can refer to the corresponding contents in the above embodiments, which will not be repeated here.

[0133] S12: Fill the object information of the target object into the first preset prompt word template to obtain a first target prompt word.

[0134] Using the above example:

[0135] For the items to be filled in for filling in the defense reasons, the order information such as the reason for the chargeback, logistics information, payment information, and product type can be filled in the corresponding preset prompt word template to obtain the prompt word. Based on the prompt word, LLM selects the target defense reason from multiple alternative defense reasons.

[0136] For the items to be filled in for the logistics order number and the tracking website, the logistics information can be filled in the corresponding preset prompt word template to obtain the prompt word. Based on the prompt word, LLM first identifies the logistics order number, then identifies the logistics company, and determines the tracking website corresponding to the logistics company according to the preset correspondence between the logistics company and the tracking website.

[0137] For the unfilled item for filling in the delivery time, the logistics information can be filled in the corresponding preset prompt word template to obtain the prompt word. The LLM model recognizes the logistics information based on the prompt word, obtains the delivery information and obtains the delivery time.

[0138] For the shipping pictures, the pictures in the communication records can be filled into the corresponding preset prompt word template to obtain the prompt word. The LLM model can identify the picture based on the prompt word to determine whether it is a shipping picture.

[0139] In the above S2, the first recommendation information may be sent to the terminal via a network. Optionally, a reason or basis for LLM to recommend the first recommendation information may also be sent to the terminal.

[0140] 207. Receive the first recommendation information returned by the server.

[0141] Exemplarily, the terminal receives the first recommendation information and the reason for recommendation or the basis for recommendation returned by the server through the network.

[0142] 208. Determine the first reply according to the first recommendation information.

[0143] Exemplarily, the first recommendation information and the reason for recommendation or the basis for recommendation may be used as the first reply.

[0144] In one embodiment, the form to be filled in requires the user to enter a summary text (referred to as summary text) to summarize the information collected by the form to be filled in. Taking the defense against payment scenario as an example, the form to be filled in requires the user to enter a summary statement for defense, and the professionalism of the summary statement has a great impact on the defense result. Generally speaking, editing the summary statement is not only difficult, but also very time-consuming. In order to improve the efficiency of inputting the summary text, the form to be filled in also includes a second item to be filled in and a third control, and the second item to be filled in is used to fill in the summary text. Among them, the summary text refers to the summary of the information collected by the form to be filled in. The above method may also include:

[0145] 209. In response to a trigger operation on the third control, a second target request is sent to the server and a second question posed to the virtual assistant is displayed on the target interface.

[0146] Among them, the second target request includes third indication information, and the third indication information is used to indicate the second item to be filled in. The second target request is used to request the server to provide second recommendation information using the large language model according to the filled-in information of the first item to be filled in, and the second recommendation information is used to fill in the second item to be filled in.

[0147] Exemplarily, the third indication information may include the name of the second item to be filled in. Exemplarily, the name of the item for filling in the closing statement is closing statement.

[0148] Exemplarily, the second target request is generated by the terminal in response to a trigger operation on the third control.

[0149] In an optional implementation manner, after receiving the second target request, the server performs the following steps:

[0150] S3. Determine second recommended information for filling in the second item to be filled in using a large language model based on the information already filled in the first item to be filled in.

[0151] S4. Return the second recommendation information to the terminal.

[0152] In the above S3, the server can obtain the filled-in information of other items to be filled in (including the first item to be filled in) except the second item to be filled in in the form to be filled in; generate a second target prompt word according to the filled-in information of other items to be filled in, and input the second target prompt word into the LLM model to obtain second recommended information. The second target prompt word is used to prompt the large language model to provide second recommended information for filling in the second item to be filled in according to the filled-in information.

[0153] A method for constructing a second target prompt word is described in detail below. The method includes the following steps:

[0154] S11. Obtain a second preset prompt word template adapted to the second item to be filled in.

[0155] In one embodiment, a second preset prompt word template adapted to the second item to be filled in may be acquired according to the correspondence between the preset items to be filled in and the preset prompt word templates.

[0156] In one embodiment, the second preset prompt word template may include a summary text example. In this way, when the model generates the summary text, it can refer to the summary text example for generation, which helps to improve the accuracy and professionalism of the generation.

[0157] In the chargeback defense scenario, the summary text corresponding to different chargeback reasons may have different wording. Therefore, different preset prompt word templates can be configured for different chargeback reasons in advance, and a correspondence between the chargeback reasons and the preset prompt templates can be established. The preset prompt word template corresponding to the chargeback reason includes a summary text example adapted to the chargeback reason. In this way, the preset prompt word template adapted to the target chargeback reason of the target order can be obtained later as the second preset prompt word template.

[0158] Of course, in actual applications, the wording of the summary texts corresponding to different banks may also be different. Therefore, for the same reason for rejection, different preset prompt word templates can be configured for different banks, and a correspondence between banks, reasons for rejection and preset prompt templates can be established. For the same reason for rejection, the summary text examples in the preset prompt word templates corresponding to different banks are different. In this way, based on the target bank involved in the target order (that is, the bank used by the buyer when paying) and the target reason for rejection of the target order, a preset prompt word template adapted to the target bank and the target reason for rejection is obtained as the second preset prompt word template. This can further improve the accuracy and professionalism of summary text generation.

[0159] S12: Fill in the filled-in information of the other items to be filled in except the second item to be filled in in the form to be filled in into the second preset prompt word template to obtain a second target prompt word.

[0160] 210. Receive the second recommendation information returned by the server.

[0161] Exemplarily, the terminal receives the second recommendation information returned by the server through the network.

[0162] 211. Displaying a second response of the virtual assistant to the second question on the target interface.

[0163] The second reply includes the second recommendation information.

[0164] Optionally, the above method may further include the following steps:

[0165] 212. Fill the second recommendation information into the input box of the second item to be filled in and display the second recommendation information in the input box.

[0166] like Figure 4 As shown, the second item to be filled in 436 includes an input box 438, and the summary text generated by the LLM model can be displayed in the input box 438.

[0167] Optionally, the above method may further include the following steps:

[0168] 213. Receive a modification operation of the user on the second recommended information in the input box.

[0169] Among them, modification operations such as wording changes, sentence structure adjustments, etc. performed by the user on the second recommendation information in the input box can be received.

[0170] 214. Modify the second recommendation information in the input box according to the modification operation.

[0171] In this embodiment, receiving a manual modification operation of the user on the summary text generated by the LLM model can further improve the professionalism of the summary text.

[0172] In the embodiment of the present application, the summary text generated by the model (ie, the second recommendation information) belongs to artificial intelligence generated content (AI Generated Content, AIGC).

[0173] Optionally, the virtual assistant can also provide consulting services to merchants, such as: consultation on chargeback notification, consultation on filling out defense materials, consultation on modifying defense materials, consultation on reasons for non-guarantee, and consultation on reasons for losing a case.

[0174] Figure 5 FIG. 1 is a flow chart of a form filling method provided by another embodiment of the present application. The execution subject of the method is the server. Figure 5 As shown, the method includes:

[0175] 501. Receive a first target request sent by a terminal.

[0176] Among them, the first target request includes first indication information and second prompt information, the first indication information is used to indicate the first item to be filled in the form to be filled in, the second indication information is used to indicate the target object, and the form to be filled in is used to collect information related to the target object.

[0177] 502. Determine, according to the target object, first recommended information for filling in the first item to be filled in by using a large language model.

[0178] 503. Return the first recommendation information to the terminal.

[0179] The specific implementation of the above steps 501 to 503 can refer to the corresponding contents in the above embodiments, which will not be repeated here.

[0180] The embodiment of the present application provides a data filling solution, which uses the LLM model to provide relevant filling recommendations for the items to be filled in the form, thereby improving the filling efficiency and reducing the time spent on filling.

[0181] In an optional implementation manner, the “using a large language model to determine first recommended information for filling in the first item to be filled in” in 502 above can be implemented by the following steps:

[0182] 5021. Obtain object information of the target object.

[0183] 5022. Obtain a first preset prompt word template that is compatible with the first item to be filled in.

[0184] 5023. Fill the object information of the target object into the first preset prompt word template to obtain a first target prompt word.

[0185] The first target prompt word is used to prompt the large language model to provide first recommended information for filling in the first item to be filled in according to the object information.

[0186] 5024. Input the first target prompt word into the large language model to obtain the first recommendation information.

[0187] The specific implementation of the above steps 5021 to 5024 can refer to the corresponding contents in the above embodiments, which will not be repeated here.

[0188] In an optional implementation, the above method may further include:

[0189] 504. Receive a second target request sent by the terminal.

[0190] The second target request includes third indication information, and the third indication information is used to indicate a second item to be filled in the form to be filled in, and the second item to be filled in is used to fill in a summary text.

[0191] 505. Obtain a second preset prompt word template adapted to the second item to be filled in.

[0192] 506. Fill in the filled information of the first item to be filled in into the second preset prompt word template to obtain a second target prompt word.

[0193] The second target prompt word is used to prompt the large language model to provide second recommended information for filling in the second item to be filled in based on the filled-in information of the first item to be filled in.

[0194] 507. Input the second target prompt word into the large language model to obtain the second recommendation information.

[0195] 508. Return the second recommendation information to the terminal.

[0196] The specific implementation of the above steps 504 to 508 can refer to the corresponding contents in the above embodiments, which will not be repeated here.

[0197] In an optional implementation, the second preset prompt word template includes a summary text example.

[0198] In an optional implementation, the target object includes a target order, the form to be filled in is used to collect rebuttal information for a chargeback application, the chargeback application is filed against the target order, and the rebuttal information is related to the target order.

[0199] In an optional implementation, the above method further includes:

[0200] 509. Obtain the defense information collected by the form to be filled in and the third preset prompt word template.

[0201] The third preset prompt word template may be obtained based on prompt word engineering optimization.

[0202] Exemplarily, the third preset prompt word template may include a preset judgment basis, and the preset judgment basis is used to determine whether the defense is successful or unsuccessful. The judgment basis can be set according to actual needs, and the embodiment of the present application does not specifically limit this.

[0203] 510. Fill the defense information into the third preset prompt word template to obtain a third target prompt word.

[0204] The third preset prompt word template is used to prompt the large language model to estimate the defense result of the defense material.

[0205] Exemplarily, the third preset prompt word template is used to prompt the large language model to estimate the defense result of the defense material according to a preset judgment basis.

[0206] 511. Input the third target prompt word into the large language model to estimate the defense result of the defense data.

[0207] In one embodiment, the large language model may generate a defense result of the defense material based on the third target prompt word using model parameters.

[0208] In another embodiment, the large language model includes an external knowledge base, and the rebuttal material in the third target prompt word can be first matched with the positive corpus and the negative corpus in the external knowledge base. If the positive corpus is matched, the rebuttal result of the rebuttal material is estimated to be a successful rebuttal; if the negative corpus is matched, the rebuttal result of the rebuttal material is estimated to be a failed rebuttal; if neither the positive corpus nor the negative corpus is matched, the model parameters are used to generate the rebuttal result of the rebuttal material.

[0209] In one embodiment, when the defense result of the defense document is a successful defense, it is determined that the target order is provided with a chargeback protection service. When the defense result of the defense document is a failed defense, it is determined that the target order is not provided with a chargeback protection service.

[0210] Optionally, in a chargeback defense scenario, a defense document (such as a word document or a PDF document) can be generated based on the defense information collected in the form, and the defense document can be subsequently submitted to the bank for defense application.

[0211] In this embodiment, the information is collected through an online form, and the collected information is structured, thus facilitating the generation of a word or pdf file of the defense.

[0212] In actual applications, the accompanying defense assistant can be used to provide merchants with filling guidance and suggestions, reducing the time spent on filling out the form from hours to minutes, effectively improving the user experience.

[0213] It should be noted here that: for the contents of each step not fully described in detail in the method provided in the embodiment of the present application, please refer to the corresponding contents in the above embodiment, which will not be repeated here. In addition, in addition to the above steps, the method provided in the embodiment of the present application may also include other parts or all of the steps in the above embodiments, which can be specifically referred to the corresponding contents in the above embodiments, which will not be repeated here.

[0214] Figure 6 A schematic diagram of a form filling system provided by an embodiment of the present application is shown. The form filling system includes: a terminal 601 and a server 602; wherein:

[0215] The terminal 601 is used to: display a target interface, the target interface includes a form to be filled in, the form to be filled in is used to collect information related to the target object, and the form to be filled in includes a first item to be filled in and a first control; in response to a trigger operation on the first control, send a first target request to the server and display a first question posed to the virtual assistant on the target interface, the first target request includes first indication information and second indication information, the first indication information is used to indicate the first item to be filled in, the second indication information is used to indicate the target object, and the first target request is used to request the server to provide the first recommendation information based on the target object using the large language model;

[0216] The server 602 is used to: receive a first target request sent by the terminal; determine the first recommendation information according to the target object using a large language model; and return the first recommendation information to the terminal.

[0217] The terminal 601 is further used to: receive the first recommendation information returned by the server; and display the first reply of the virtual assistant to the first question on the target interface, wherein the first reply includes the first recommendation information.

[0218] The embodiment of the present application provides a companion data filling solution, that is, when the user is filling in the form, a virtual assistant is used to provide relevant filling recommendations for the items to be filled in the form, thereby improving the filling efficiency and reducing the time spent on filling.

[0219] It should be noted here that the specific implementation of the terminal, the server and the interaction process between the two provided in the embodiments of the present application can be found in the corresponding contents of the above embodiments, which will not be repeated here.

[0220] Terminal: can be a mobile terminal, a fixed terminal or a portable terminal, such as a mobile phone, a station, a unit, a device, a multimedia computer, a multimedia tablet, an Internet node, a communicator, a desktop computer, a laptop computer, a notebook computer, a netbook computer, a tablet computer, a personal communication system device, a personal navigation device, a personal digital assistant, an audio / video player, a digital camera / camcorder, a positioning device, a television receiver, a radio broadcast receiver, an electronic book device, a gaming device or any combination thereof, including accessories, peripherals or any combination thereof of these devices. It is also foreseeable that the terminal device can support any type of interface device for the user (such as a wearable device), etc.

[0221] Server: A server can be one or more servers. A server can also be a physical server or a virtual server. A server can be an independent physical server, a server cluster or a distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, and big data and artificial intelligence platforms.

[0222] like Figure 7 As shown, the form filling solution provided in the embodiment of the present application provides rich technical expression capabilities with the help of technical base capabilities such as database query capabilities, LLM model generation, dialogue, multimodal understanding and agent, among which the technical expression capabilities include: structured form filling and recommendation, defense reason generation, summary statement generation and voucher screenshot generation and understanding. Through this technical expression capability, fill-in recommendation information is generated or obtained, such as: payment voucher, logistics voucher, defense reason and summary statement.

[0223] Figure 8 FIG. 1 is a schematic diagram showing the structure of an electronic device provided by an embodiment of the present application. Figure 8As shown, the electronic device includes a memory 1101 and a processor 1102. The memory 1101 can be configured to store various other data to support operations on the electronic device. Examples of these data include instructions for any application or method for operating on the electronic device. The memory 1101 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as a static random access memory (SRAM), an electrically erasable programmable read only memory (EEPROM), an erasable programmable read only memory (EPROM), a programmable read only memory (PROM), a read only memory (ROM), a magnetic memory, a flash memory, a magnetic disk or an optical disk.

[0224] The memory 1101 is used to store programs;

[0225] The processor 1102 is coupled to the memory 1101 and is used to execute the program stored in the memory 1101 to implement the methods provided by the above-mentioned method embodiments.

[0226] Further, if Figure 8 As shown, the electronic device also includes: a communication component 1103, a display 1104, a power component 1105, an audio component 1106 and other components. Figure 8 Only some components are shown schematically, which does not mean that the electronic device only includes Figure 8 Components shown.

[0227] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a computer, the steps or functions of the methods provided in the above-mentioned method embodiments can be implemented.

[0228] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, can implement the steps or functions of the methods provided in the above-mentioned method embodiments.

[0229] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0230] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM (Read Only Memory) / RAM (Random Access Memory), a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0231] Having thus described several aspects and embodiments of the technology of the present application, it should be understood that various changes, modifications and improvements will be readily conceived by those of ordinary skill in the art. These changes, modifications and improvements are intended to fall within the spirit and scope of the technology described in the present application. Therefore, it should be understood that the foregoing embodiments are presented only by way of example, and within the scope of the appended claims and their equivalents, embodiments of the present invention may be practiced in a manner different from that specifically described. In addition, if two or more features, systems, articles, materials and / or methods described herein are not mutually contradictory, any combination of these features, systems, articles, materials and / or methods is included within the scope of the present disclosure. Moreover, as described above, some aspects may be embodied as one or more methods. The actions performed as part of the method may be ordered in any suitable manner. Therefore, an embodiment in which actions are performed in an order different from the order shown may be constructed, which may include performing some actions simultaneously, even if shown as sequential actions in the illustrative embodiments.

Claims

1. A form filling method, applicable to a terminal, characterized in that: include: Displaying a target interface, wherein the target interface includes a form to be filled in, and the form to be filled in includes a first item to be filled in and a first control; In response to a trigger operation on the first control, displaying a first question posed to the virtual assistant on the target interface; The first response of the virtual assistant to the first question is displayed on the target interface, wherein the first response includes first recommended information for filling in the first item to be filled in.

2. The method according to claim 1, characterized in that Also includes: Displaying a second control on the target interface; In response to a triggering operation on the second control, the first recommendation information is filled into the first item to be filled in.

3. The method according to claim 2, characterized in that Displaying a second control on the target interface includes: Displaying the first recommendation information and the second control in a preset area of ​​the target interface; Among them, in the target interface, the preset area is associated with the area where the first item to be filled in is located.

4. The method according to any one of claims 1 to 3, characterized in that The form to be filled in is used to collect information related to the target object.

5. The method according to claim 4, characterized in that The virtual assistant is implemented based on a large language model running on the server; the method further includes: In response to a trigger operation on the first control, a first target request is sent to the server, wherein the first target request includes first indication information and second indication information, the first indication information is used to indicate the first item to be filled in, the second indication information is used to indicate the target object, and the first target request is used to request the server to provide the first recommendation information according to the target object by using the large language model; Receiving the first recommendation information returned by the server; The first reply is determined according to the first recommendation information.

6. The method according to claim 4, characterized in that The form to be filled in includes a second item to be filled in and a third control, wherein the second item to be filled in is used to fill in the summary text; In response to a trigger operation on the third control, a second target request is sent to the server and a second question posed to the virtual assistant is displayed on the target interface, wherein the second target request includes third indication information, the third indication information is used to indicate the second item to be filled in, the second target request is used to request the server to provide second recommendation information using the large language model according to the filled-in information of the first item to be filled in, and the second recommendation information is used to fill in the second item to be filled in; receiving the second recommendation information returned by the server; The second response of the virtual assistant to the second question is displayed on the target interface, wherein the second response includes the second recommendation information.

7. The method according to claim 4, characterized in that The target object includes a target order, and the form to be filled in is used to collect defense information for a chargeback application, and the chargeback application is filed for the target order.

8. The method according to any one of claims 1 to 3, characterized in that The first reply also includes the recommendation basis or recommendation reason corresponding to the first recommendation information.

9. The method according to claim 8, characterized in that Also includes: When the first reply includes a recommendation basis corresponding to the first recommendation information, displaying a fourth control on the target interface; In response to a trigger operation on the fourth control, a detail interface is displayed, wherein the detail interface is used to display detailed information about the recommendation basis.

10. A form filling method, applicable to a server, characterized in that: include: Receiving a first target request sent by a terminal, the first target request including first indication information and second prompt information, the first indication information is used to indicate a first to-be-filled item of a to-be-filled form, the second indication information is used to indicate a target object, and the to-be-filled form is used to collect information related to the target object; Determining, according to the target object, first recommended information for filling in the first item to be filled in by using a large language model; The first recommendation information is returned to the terminal.

11. The method according to claim 10, characterized in that Determining, according to the target object, first recommended information for filling in the first item to be filled in by using a large language model, includes: Acquire object information of the target object; Acquire a first preset prompt word template adapted to the first item to be filled in; Filling the object information of the target object into the first preset prompt word template to obtain a first target prompt word, wherein the first target prompt word is used to prompt the large language model to provide first recommended information for filling in the first to-be-filled item according to the object information; The first target prompt word is input into the large language model to obtain the first recommendation information.

12. The method according to claim 10, characterized in that Also includes: receiving a second target request sent by the terminal, wherein the second target request includes third indication information, wherein the third indication information is used to indicate a second item to be filled in the form to be filled in, and the second item to be filled in is used to fill in a summary text; Acquire a second preset prompt word template adapted to the second item to be filled in; Filling the filled-in information of the first item to be filled in into the second preset prompt word template to obtain a second target prompt word, wherein the second target prompt word is used to prompt the large language model to provide second recommended information for filling in the second item to be filled in according to the filled-in information of the first item to be filled in; Inputting the second target prompt word into the large language model to obtain the second recommendation information; The second recommendation information is returned to the terminal.

13. The method according to claim 12, characterized in that The second preset prompt word template includes a summary text example.

14. The method according to any one of claims 10 to 13, characterized in that The target object includes a target order, the form to be filled in is used to collect defense materials for a rejection application, the rejection application is filed for the target order, and the defense materials are related to the target order.

15. The method according to claim 14, characterized in that Also includes: Obtaining the defense information collected by the form to be filled out and the third preset prompt word template; Filling the objection data into the third preset prompt word template to obtain a third target prompt word, wherein the third preset prompt word template is used to prompt the large language model to estimate the objection result of the objection data; The third target prompt word is input into the large language model to estimate the defense result of the defense material.

16. A form filling system, characterized in that: include: Terminal and server; among them, The terminal is used to: display a target interface, the target interface includes a form to be filled in, the form to be filled in is used to collect information related to a target object, and the form to be filled in includes a first item to be filled in and a first control; in response to a trigger operation on the first control, send a first target request to the server and display a first question posed to the virtual assistant on the target interface, the first target request includes first indication information and second indication information, the first indication information is used to indicate the first item to be filled in, the second indication information is used to indicate the target object, and the first target request is used to request the server to provide the first recommendation information based on the target object using the large language model; The server is configured to: receive a first target request sent by the terminal; determine the first recommendation information using a large language model according to the target object; and return the first recommendation information to the terminal; The terminal is further used to: receive the first recommendation information returned by the server; and display the first reply of the virtual assistant to the first question on the target interface, wherein the first reply includes the first recommendation information.

17. An electronic device, characterized in that: include: A memory and a processor, wherein The memory is used to store programs; The processor is coupled to the memory, and is used to execute the program stored in the memory to implement the method of any one of claims 1 to 9, or to implement the method of any one of claims 10 to 15.

18. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a computer, the method according to any one of claims 1 to 9 can be implemented, or the method according to any one of claims 10 to 15 can be implemented.

19. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 9 is implemented, or the method according to any one of claims 10 to 15 is implemented.

Images