Output result determination method and device, client, server and storage medium

Abstract

The embodiment of the invention provides an output result determination method and device, a client side, a server side and a storage medium, and relates to the technical field of retrieval enhancement generation. The method comprises the steps of obtaining input information, obtaining a corresponding text in a preset text address, and generating a first prompt word according to the text and the input information; searching at least one text fragment associated with the first cue word in a preset knowledge base, traversing in a preset index relationship library according to the at least one text fragment to obtain a first traversal result, and generating a first text set according to the first traversal result; and generating a second cue word according to the input information, the first text set and a preset cue word template, inputting the second cue word into a preset large language model, and determining an output result according to the second cue word based on the preset large language model. By the adoption of the technical scheme, the accuracy of the output result can be improved in the question and answer retrieval process based on the multi-document content.

Description

Technical Field

[0001] The embodiments of the present disclosure relate to the field of search enhancement generation technology, and in particular to an output result determination method, device, client, server, and storage medium. Background Art

[0002] Current large language models can combine external knowledge bases in natural language processing tasks and answer user input questions through retrieval and generation technologies.

[0003] However, in a retrieval question-answering system based on multi-document content, the inconsistency of information between documents and the lack of structure in the content of different documents result in a relatively low accuracy of the final output results.

[0004] Therefore, there is an urgent need for an output result determination method that can improve the accuracy of output results in the retrieval question answering process based on multi-document content. Summary of the Invention

[0005] In order to solve the above technical problems or at least partially solve the above technical problems, the embodiments of the present disclosure provide an output result determination method, device, client, server and storage medium.

[0006] A first aspect of an embodiment of the present disclosure provides a method for determining an output result, the method comprising:

[0007] Obtaining input information, obtaining corresponding text from a preset text address, and generating a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent content to be queried; and the first prompt word is used to query the preset knowledge base;

[0008] Searching a preset knowledge base for at least one text segment associated with the first prompt word, traversing a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generating a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between the text segment and other text segments;

[0009] A second prompt word is generated based on the input information, the first text set, and a preset prompt word template, the second prompt word is input into a preset large language model, and an output result is determined based on the second prompt word and the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

[0010] In one example, obtaining corresponding text from a preset text address and generating a first prompt word according to the text and the input information includes:

[0011] Obtaining corresponding text from a preset text address, and obtaining page configuration information corresponding to the text; wherein the page configuration information is pre-configured in the text;

[0012] Based on the preset large language model, extracting data structure information and configuration rules of the text according to the page configuration information;

[0013] The first prompt word is generated according to the input information, the data structure information and the configuration rule.

[0014] In one example, generating the first prompt word according to the input information, the data structure information, and the configuration rule includes:

[0015] Based on the preset large language model, extracting field information from the data structure information according to the input information;

[0016] Searching the configuration rule for text description information associated with the field information, and determining the text description information as the first prompt word; wherein the configuration rule is used to characterize a mapping relationship between the field information and the text description information.

[0017] In one example, traversing a preset index relationship library according to the at least one text segment to obtain a first traversal result, and generating a first text set according to the first traversal result includes:

[0018] Traversing each of the text segments in the preset index relationship library to obtain a first traversal result;

[0019] If the first traversal result indicates that there are other text addresses associated with the text segment, the text segment and the other text addresses associated with the text segment are used to generate a target text segment; wherein the other text addresses associated with the text segment are used to represent the addresses of other texts linked to the text segment;

[0020] If the first traversal result indicates that there is no other text address associated with the text segment, generating a target text segment from the text segment;

[0021] The first text set is generated by using at least one of the target text segments.

[0022] In one example, after generating the first text set from at least one target text segment, the method further includes:

[0023] Calculating the similarity between the first prompt word and each target text segment in the first text set based on a preset similarity algorithm;

[0024] The target text segment whose similarity satisfies a first preset condition is determined as the first target text.

[0025] In one example, determining an output result based on the second prompt word based on the preset large language model includes:

[0026] Outputting an initial result based on the preset large language model according to the second prompt word; wherein the initial result is used to represent a result initially obtained by the preset large language model according to the second prompt word;

[0027] The output result is determined based on the preset large language model according to the matching degree between the initial result and the input information; wherein the output result is used to represent the target result determined based on the preset large language model.

[0028] In one example, determining the output result based on the preset large language model according to the matching degree between the initial result and the input information includes:

[0029] If the matching degree satisfies a second preset condition, the initial result is determined as the output result; wherein the second preset condition is that the matching degree is greater than a threshold;

[0030] If the matching degree does not meet the second preset condition, other text addresses associated with the text fragment in the first text set are queried, and the output result is determined based on the other text addresses; wherein the other text addresses associated with the text fragment are used to represent the addresses of other texts linked to in the text fragment.

[0031] In one example, determining the output result according to the other text address includes:

[0032] Obtaining corresponding other text from the other text address, and generating a third prompt word based on the other text and the input information; wherein the other text address is used to access the other text; and the third prompt word is used to search in the preset knowledge base;

[0033] Searching a preset knowledge base for at least one text segment associated with the third prompt word, traversing a preset index relationship library based on the at least one text segment to obtain a second traversal result, and generating a second text set based on the second traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between the text segment and other text segments;

[0034] A fourth prompt word is generated based on the input information, the second text set, and a preset prompt word template, the fourth prompt word is input into a preset large language model, and an output result is determined based on the fourth prompt word based on the preset large language model; wherein the output result is used to be displayed on a terminal device.

[0035] In one example, the text segment is generated from a plurality of original texts.

[0036] In one example, the method further includes:

[0037] Extracting data structure information from the original text based on the preset large language model;

[0038] Extracting preset field information from the data structure information based on the preset large language model;

[0039] The preset field information and the preset text description information are associated to obtain an association relationship, and page configuration information is generated according to the association relationship.

[0040] In one example, the method further includes:

[0041] Segmenting the original text based on the preset large language model to obtain at least one text segment;

[0042] Based on the preset large language model, keywords in the text segment are identified, and the keywords and the at least one text segment are combined to generate the preset knowledge base.

[0043] In one example, the method further includes:

[0044] Based on the preset large language model, if it is detected that other text addresses exist in the text fragment, the association relationship between the text fragment and the other text addresses in the text fragment is generated into the preset index relationship library; wherein the other text address is used to represent the address of the text linked to in the text fragment.

[0045] A second aspect of an embodiment of the present disclosure provides an output result determination device, the device comprising:

[0046] A first acquisition module is configured to acquire input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to perform a query in the preset knowledge base;

[0047] A first search module is configured to search a preset knowledge base for at least one text segment associated with the first prompt word, perform a traversal in a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between the text segment and other text segments;

[0048] A first determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

[0049] A third aspect of the embodiments of the present disclosure provides a client, including:

[0050] a second acquisition module, configured to acquire local input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to perform a query in the preset knowledge base;

[0051] a second search module configured to search a preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between the text segment and other text segments;

[0052] A second determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

[0053] A fourth aspect of the embodiments of the present disclosure provides a server, including:

[0054] a third acquisition module, configured to acquire input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; the first prompt word is used to query the preset knowledge base; wherein the input information is sent by the client to the server;

[0055] a third search module configured to search a preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between the text segment and other text segments;

[0056] A third determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

[0057] A fifth aspect of the embodiments of the present disclosure provides a terminal device, including:

[0058] A processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the method described in the first aspect.

[0059] A sixth aspect of an embodiment of the present disclosure provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the method of the first aspect described above can be implemented.

[0060] The present disclosure provides a method, device, client, server, and storage medium for determining an output result. The method comprises: obtaining input information, obtaining corresponding text from a preset text address, and generating a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; the first prompt word is used to query the preset knowledge base; searching the preset knowledge base for at least one text segment associated with the first prompt word, and traversing a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generating a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; the preset index relationship library is used to store the association relationship between the text segment and other text segments; generating a second prompt word based on the input information, the first text set, and a preset prompt word template; inputting the second prompt word into a preset large language model; and determining an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result. The present technical solution can improve the accuracy of output results in the retrieval question-answering process based on multi-document content. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0062] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0063] Figure 1 is a flowchart of a method for determining an output result provided by an embodiment of the present disclosure;

[0064] Figure 2a is a flowchart of a method for determining an output result provided by an embodiment of the present disclosure;

[0065] Figure 2b is a schematic diagram of a process for screening target text segments provided by an embodiment of the present disclosure;

[0066] Figure 2c This is a schematic diagram of an overall structure for determining an output result provided by an embodiment of the present disclosure;

[0067] Figure 2dThis is a schematic diagram of a process for determining an output result provided by an embodiment of the present disclosure;

[0068] Figure 3 is a flowchart of a method for determining an output result provided by an embodiment of the present disclosure;

[0069] Figure 4 This is a schematic diagram of a process for offline processing of original text provided by an embodiment of the present disclosure;

[0070] Figure 5 is a structural diagram of an output result determination device provided by an embodiment of the present disclosure;

[0071] Figure 6 It is a structural diagram of a terminal device in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0072] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features therein can be combined with each other in the absence of conflict.

[0073] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0074] Figure 1 This is a flow chart of a method for determining an output result provided by an embodiment of the present disclosure, which can be executed by a terminal device. The terminal device can be exemplarily understood as a device such as a mobile phone, tablet computer, laptop computer, desktop computer, smart TV, etc. Figure 1 As shown, the method provided in this embodiment includes the following steps:

[0075] S101. Obtain input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to query in a preset knowledge base.

[0076] In one example, the input information is in natural language, for example, a query about the meaning of an error code. Specifically, the input information may be "What does 47-1-510-0 mean?" The preset text address may be a default text address from which text can be retrieved. The text represents the content to be queried. The first prompt word is used to query a preset knowledge base.

[0077] S102: Searching for at least one text segment associated with the first prompt word in a preset knowledge base, traversing a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generating a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store association relationships between text segments and other text segments.

[0078] In one example, a preset knowledge base includes multiple pre-stored text segments. The text segments are generated offline from multiple original texts. Specifically, each original text is segmented offline to obtain multiple text segments, which are then stored in the preset knowledge base. This allows the preset knowledge base to store multiple original texts, allowing searches to be performed across the multiple original texts during queries.

[0079] In this embodiment, a preset index relationship library is used to store the associations between text segments and other text segments. For example, a preset knowledge base is searched for at least one text segment associated with a first prompt word, namely, text segment A, text segment B, and text segment C. Then, a traversal is performed in the preset index relationship library according to text segments A, B, and C to obtain a first traversal result. A first text set is then generated based on the first traversal result. The first text set includes text segments A, B, and C.

[0080] S103: Generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate an output result.

[0081] In one example, the preset prompt word template is preconfigured. The preset prompt word template is used to fill in the input information and the first text set, and then obtain the second prompt word. The second prompt word is input into the preset large language model, and the preset large language model is instructed to determine an output result based on the second prompt word, and the output result is displayed on the terminal device.

[0082] The disclosed embodiment provides a method for determining an output result, the method comprising: obtaining input information, obtaining corresponding text in a preset text address, generating a first prompt word based on the text and the input information, then searching for at least one text segment associated with the first prompt word in a preset knowledge base, and traversing a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generating a first text set based on the first traversal result, and finally generating a second prompt word based on the input information, the first text set, and a preset prompt word template, inputting the second prompt word into a preset large language model, and determining an output result based on the second prompt word based on the preset large language model. By adopting this technical solution, since the preset knowledge base stores multiple text segments, and the preset index relationship library stores other text segments associated with the text segments, the accuracy of the output results can be improved in the retrieval question-answering process based on multi-document content.

[0083] Figure 2a The following is a flow chart of a method for determining an output result provided by an embodiment of the present disclosure. The present disclosure embodiment is optimized based on the above embodiment, and the present disclosure embodiment can be combined with various optional solutions in one or more of the above embodiments.

[0084] like Figure 2a As shown, the output result determination method may include the following steps:

[0085] S201. Obtain input information, obtain corresponding text in a preset text address, and obtain page configuration information corresponding to the text; wherein the page configuration information is pre-configured in the text; wherein the preset text address is used to access the text; and the text is used to represent the content to be queried.

[0086] In one example, the page configuration information includes the data structure information and configuration rules of the text. In this embodiment, different texts correspond to different page configuration information. The page configuration information may be Query-Rewrite configuration information.

[0087] S202: Based on the preset large language model, extract the data structure information and configuration rules of the text according to the page configuration information.

[0088] In an example, the data structure information of the text may be Page Schema data structure information, and the configuration rule may be an association relationship between preset field information and preset text description information.

[0089] In one example, based on a preset large language model, the data structure information and configuration rules of the text are extracted according to the page configuration information. For example, the data structure information of the extracted text is as follows:

[0090]

[0091] As can be seen above, the data structure of the text is in JSON format. The configuration rule can be that 1 in business indicates a client error.

[0092] S203: Generate a first prompt word according to the input information, data structure information and configuration rules.

[0093] In one example, the first prompt word is used to search in a preset knowledge base.

[0094] In an example, the input information is "What does error code 1-2-3 mean?", the data structure information includes business=1, and the configuration rule is 1 is client error, then the first prompt word is client error.

[0095] In one example, generating a first prompt word according to input information, data structure information, and configuration rules includes:

[0096] Based on the preset large language model, field information is extracted from the data structure information according to the input information;

[0097] The configuration rule is searched for text description information associated with the field information, and the text description information is determined as the first prompt word; wherein the configuration rule is used to represent the mapping relationship between the field information and the text description information.

[0098] In one example, the input information is "What does error code 1-2-3 mean?" "Error code 1" is first extracted from the data structure information. The field information may be business = 1. The text description information associated with the field information in the configuration rule refers to the textual explanation corresponding to the field information. For example, if the text description of "1" in business = 1 in the configuration rule is client error, then business = 1 in the data structure information is replaced with client error, and client error is used as the first prompt word.

[0099] S204: Searching for at least one text segment associated with the first prompt word in a preset knowledge base, traversing a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generating a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between text segments and other text segments.

[0100] In one example, traversing a preset index relationship library according to at least one text segment to obtain a first traversal result, and generating a first text set according to the first traversal result includes:

[0101] Traverse each text segment in the preset index relationship library to obtain the first traversal result;

[0102] If the first traversal result indicates that there are other text addresses associated with the text segment, the text segment and the other text addresses associated with the text segment are used to generate a target text segment; wherein the other text addresses associated with the text segment are used to represent the addresses of other texts linked to the text segment;

[0103] If the first traversal result indicates that there is no other text address associated with the text segment, the text segment is generated into a target text segment;

[0104] A first text set is generated from at least one target text segment.

[0105] In one example, for example, a text segment A is traversed in a preset index relationship library to obtain a first traversal result. If the first traversal result indicates that there is another text address D associated with the text segment A, the text segment A and the other text address D associated with the text segment A are combined to generate a target text segment E.

[0106] In one example, if the first traversal result indicates that there are other text addresses F associated with the text segment B, the text segment B and the other text addresses F associated with the text segment B are generated into a target text segment G.

[0107] In one example, if the first traversal result indicates that there is no other text address associated with the text segment C, that is, the text segment C is the last level text, then the text segment C is generated into a target text segment H.

[0108] In one example, the first text set includes a target text segment E, a target text segment G, and a target text segment H.

[0109] In one example, after generating a first text set from at least one target text segment, the method further includes:

[0110] Calculating the similarity between the first prompt word and each target text segment in the first text set based on a preset similarity algorithm;

[0111] The target text segment whose similarity meets the first preset condition is determined as the first target text.

[0112] In one example, the preset similarity algorithm may be a cosine similarity algorithm. The similarities between the first prompt word "client error" and the target text segment E, target text segment G, and target text segment H are calculated. For example, the similarity between the first prompt word "client error" and the target text segment E is 80%, the similarity between the first prompt word "client error" and the target text segment G is 70%, and the similarity between the first prompt word "client error" and the target text segment H is 75%.

[0113] In one example, the first preset condition can be greater than a similarity threshold, or can be the top two target text segments in the ranking. For example, if the first preset condition is a target text segment greater than a similarity threshold of 75%, then the first target text is target text segment E and target text segment H. In this embodiment, see Figure 2b A schematic diagram of a process for screening target text segments is shown.

[0114] S205: Generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and output an initial result based on the second prompt word based on the preset large language model; wherein the initial result is used to represent the result initially obtained by the preset large language model based on the second prompt word.

[0115] In one example, the second prompt word is used to instruct a preset large language model to generate an output result.

[0116] In one example, the initial result is a result initially obtained by a preset large language model based on the second prompt word, that is, the text extracted by the preset large language model from the first target text is used as the initial result.

[0117] S206 : Determine an output result based on the preset large language model according to the matching degree between the initial result and the input information; wherein the output result is used to represent the target result determined based on the preset large language model.

[0118] In one example, the matching degree is used to represent the degree of similarity between the initial result and the input information, and the output result is determined based on the similarity degree based on the preset large language model.

[0119] In one example, the output result is determined based on the matching degree between the initial result and the input information based on the preset large language model, including:

[0120] If the matching degree satisfies a second preset condition, the initial result is determined as the output result; wherein the second preset condition is that the matching degree is greater than a threshold;

[0121] If the matching degree does not meet the second preset condition, other text addresses associated with the text fragment in the first text set are queried, and the output result is determined based on the other text addresses; wherein the other text addresses associated with the text fragment are used to represent the addresses of other texts linked to in the text fragment.

[0122] In one example, the match between the initial result and the input information is determined based on a preset large language model. Specifically, the preset large language model is used to determine whether the initial result meets the user's needs. If the match meets a second preset condition, the initial result meets the user's needs, i.e., it can answer the input information, and the initial result is determined as the output result.

[0123] In one example, if the degree of match does not meet the second preset condition, it means that the initial result is not what the user needs, that is, it cannot answer the input information, and it is necessary to query other text addresses associated with the text fragment in the first text set and determine the output result based on the other text addresses; wherein, the other text addresses associated with the text fragment are used to represent the addresses of other texts linked to in the text fragment.

[0124] In one example, output results are determined based on other text addresses, including:

[0125] Obtaining corresponding other text from the other text address and generating a third prompt word based on the other text and the input information; wherein the other text address is used to access the other text; and the third prompt word is used to search in a preset knowledge base;

[0126] Searching a preset knowledge base for at least one text segment associated with the third prompt word, traversing a preset index relationship library based on the at least one text segment to obtain a second traversal result, and generating a second text set based on the second traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between text segments and other text segments;

[0127] A fourth prompt word is generated based on the input information, the second text set, and a preset prompt word template, the fourth prompt word is input into a preset large language model, and an output result is determined based on the fourth prompt word based on the preset large language model; wherein the output result is used to be displayed on a terminal device.

[0128] In one example, the process is to repeatedly execute the contents of steps S101 to S103, wherein the preset text address in step S101 is modified to another text address. For detailed description, please refer to the description of steps S101 to S103, which will not be repeated here.

[0129] To make the explanation clearer, an example of the initial result is given. For example, the initial result is in the form of:

[0130] {"action":final_answer","input":"$final_answer"}; "action" indicates the initial result, "final_answer" indicates the content of the initial result, "input" indicates the input information, and "$final_answer" is a placeholder. This format indicates that the initial result is determined as the output result.

[0131] The initial results are of the form:

[0132] {"action":"search","input":"$keyword"}; "action" indicates the initial results are displayed, "search" indicates that the initial results are not what the user needs and a new search is required. "input" indicates input information, and "$keyword" indicates the identifier corresponding to another text address. This format indicates that other text addresses need to be used to access other text and then re-determine the output results.

[0133] For a clearer explanation, see Figure 2c A schematic diagram of the overall structure for determining an output result is shown. Figure 2c The Generator part is step S206, the Retriever part is the content of steps S201-S204, and the Grader part is the content of determining the first target text.

[0134] Further, in Figure 2c By refining each step based on Figure 2d A schematic diagram of a process for determining an output result is shown.

[0135] The disclosed embodiments provide a method for determining an output result, comprising: obtaining input information, obtaining corresponding text from a preset text address, and obtaining page configuration information corresponding to the text; extracting data structure information and configuration rules of the text based on the page configuration information based on a preset large language model; then generating a first prompt word based on the input information, the data structure information, and the configuration rules; searching a preset knowledge base for at least one text segment associated with the first prompt word; traversing a preset index relationship base based on the at least one text segment to obtain a first traversal result; and generating a first text set based on the first traversal result; generating a second prompt word based on the input information, the first text set, and a preset prompt word template; inputting the second prompt word into a preset large language model; and outputting an initial result based on the second prompt word based on the preset large language model. Determining an output result based on the preset large language model based on the degree of match between the initial result and the input information; wherein the output result is used to represent a target result determined based on the preset large language model. By employing this technical solution, by rewriting field information based on the semantics of the input information and then performing a search based on the rewritten field information, it is possible to extract valid information from a cluttered document, thereby improving matching accuracy.

[0136] Figure 3 The following is a flow chart of a method for determining an output result provided by an embodiment of the present disclosure. The present disclosure embodiment is optimized based on the above embodiment, and the present disclosure embodiment can be combined with various optional solutions in one or more of the above embodiments.

[0137] like Figure 3 As shown, the output result determination method may include the following steps:

[0138] S301: Extract data structure information from the original text based on a preset large language model.

[0139] In one example, the original text is a pre-provided text to be queried. There are multiple original texts, and the original texts may be multiple unrelated texts. The structure of each original text may be different. The data structure information is Page Schema.

[0140] In one example, a text segment is generated from a plurality of original texts.

[0141] S302: Extracting preset field information from the data structure information based on a preset large language model.

[0142] In one example, based on a preset large language model, each preset field information in the data structure information is traversed and extracted. For example, the preset field information may be "business=1".

[0143] S303: Associating the preset field information with the preset text description information to obtain an association relationship, and generating page configuration information according to the association relationship.

[0144] In one example, the extracted preset field information and the preset text description information are associated. For example, if the preset field information is "business=1" and the preset text description information is "client error", business=1 is associated with the client error to obtain an association relationship and generate the final page configuration information.

[0145] S304: Segment the original text based on a preset large language model to obtain at least one text segment.

[0146] In one example, based on a preset large language model, each original text is segmented. Specifically, the segmentation can be performed according to the paragraphs, icons, or text length of the original text. If the text length exceeds the text length threshold after segmentation according to the paragraphs or icons of the original text, the segmentation can be performed again according to the text length threshold. For example, the original text is original text A, and then original text A is segmented into text segment A, text segment B, and text segment C, which can also be recorded as ChunkA, ChunkB, and ChunkC.

[0147] S305: Based on a preset large language model, identify keywords in the text segment, and generate a preset knowledge base with the keywords and at least one text segment.

[0148] In one example, based on a preset large language model, keywords in Chunk A, keywords in Chunk B, and keywords in Chunk C are identified respectively, and then the keywords in Chunk A, keywords in Chunk B, keywords in Chunk C, Chunk A, Chunk B, and Chunk C are used to generate a preset knowledge base.

[0149] S306. Based on the preset large language model, if it is detected that other text addresses exist in the text segment, a preset index relationship library is generated based on the association relationship between the text segment and the other text addresses in the text segment; wherein the other text addresses are used to represent the addresses of the texts linked to in the text segment.

[0150] In one example, the other text address can be a markdown link or an HTML hyperlink. If it is detected that there is another text address in the text segment, a corresponding association relationship is generated between the text segment and the other text address in the text segment, and then the association relationship, the text segment and the other text address in the text segment are stored in a preset index relationship library. For a clearer explanation, please refer to Figure 4 A schematic diagram of the process of offline processing of raw text is shown.

[0151] The embodiment of the present disclosure provides a method for determining an output result, the method comprising: extracting data structure information from the original text based on a preset large language model, extracting preset field information from the data structure information based on the preset large language model, associating the preset field information with preset text description information to obtain an association relationship, and generating page configuration information based on the association relationship. Based on the preset large language model, the original text is segmented to obtain at least one text segment. Based on the preset large language model, keywords in the text segment are identified, and a preset knowledge base is generated from the keywords and at least one text segment. Based on the preset large language model, if other text addresses are detected in the text segment, the association relationship between the text segment and the other text addresses in the text segment is generated into a preset index relationship library. By adopting the present technical solution, the association relationship between texts can be further extracted by offline processing of the original text, and then more association information can be extracted using the preset large language model to assist the retrieval process and improve the efficiency and accuracy of the retrieval.

[0152] Figure 5 This is a schematic diagram of the structure of an output result determination device provided by an embodiment of the present disclosure. The output result determination device can be understood as the above-mentioned terminal device or part of the functional modules in the above-mentioned terminal device. Figure 5 As shown, the output result determination device 50 includes:

[0153] The first acquisition module 501 is used to obtain input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to query in a preset knowledge base.

[0154] The first search module 502 is used to search for at least one text segment associated with the first prompt word in a preset knowledge base, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store the association relationship between the text segment and other text segments.

[0155] The first determination module 503 is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate an output result.

[0156] In one example, the first acquisition module 501 is specifically configured to:

[0157] Obtaining the corresponding text from a preset text address and obtaining page configuration information corresponding to the text; wherein the page configuration information is pre-configured in the text;

[0158] Based on the preset large language model, the data structure information and configuration rules of the text are extracted according to the page configuration information;

[0159] A first prompt word is generated according to input information, data structure information and configuration rules.

[0160] In one example, the first acquisition module 501 is specifically configured to:

[0161] Based on the preset large language model, field information is extracted from the data structure information according to the input information;

[0162] The configuration rule is searched for text description information associated with the field information, and the text description information is determined as the first prompt word; wherein the configuration rule is used to represent the mapping relationship between the field information and the text description information.

[0163] In one example, the first search module 502 is specifically configured to:

[0164] Traverse each text segment in the preset index relationship library to obtain the first traversal result;

[0165] If the first traversal result indicates that there are other text addresses associated with the text segment, the text segment and the other text addresses associated with the text segment are used to generate a target text segment; wherein the other text addresses associated with the text segment are used to represent the addresses of other texts linked to the text segment;

[0166] If the first traversal result indicates that there is no other text address associated with the text segment, the text segment is generated into a target text segment;

[0167] A first text set is generated from at least one target text segment.

[0168] In one example, the apparatus 50 further includes:

[0169] A calculation module 504 is configured to calculate the similarity between the first prompt word and each target text segment in the first text set based on a preset similarity algorithm;

[0170] The fourth determining module 505 is configured to determine the target text segment whose similarity satisfies the first preset condition as the first target text.

[0171] In one example, the first determining module 503 is specifically configured to:

[0172] Outputting an initial result based on the second prompt word based on the preset large language model; wherein the initial result is used to represent the result initially obtained by the preset large language model based on the second prompt word;

[0173] An output result is determined based on a preset large language model according to a degree of matching between the initial result and the input information; wherein the output result is used to represent a target result determined based on the preset large language model.

[0174] In one example, the first determining module 503 is specifically configured to:

[0175] If the matching degree satisfies a second preset condition, the initial result is determined as the output result; wherein the second preset condition is that the matching degree is greater than a threshold;

[0176] If the matching degree does not meet the second preset condition, other text addresses associated with the text fragment in the first text set are queried, and the output result is determined based on the other text addresses; wherein the other text addresses associated with the text fragment are used to represent the addresses of other texts linked to in the text fragment.

[0177] In one example, the first determining module 503 is specifically configured to:

[0178] Obtaining corresponding other text from the other text address and generating a third prompt word based on the other text and the input information; wherein the other text address is used to access the other text; and the third prompt word is used to search in a preset knowledge base;

[0179] Searching a preset knowledge base for at least one text segment associated with the third prompt word, traversing a preset index relationship library based on the at least one text segment to obtain a second traversal result, and generating a second text set based on the second traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between text segments and other text segments;

[0180] A fourth prompt word is generated based on the input information, the second text set, and a preset prompt word template, the fourth prompt word is input into a preset large language model, and an output result is determined based on the fourth prompt word based on the preset large language model; wherein the output result is used to be displayed on a terminal device.

[0181] In one example, a text segment is generated from a plurality of original texts.

[0182] In one example, the apparatus 50 further includes:

[0183] A first extraction module 506 is used to extract data structure information from the original text based on a preset large language model;

[0184] A second extraction module 507 is configured to extract preset field information from the data structure information based on a preset large language model;

[0185] The association module 508 is used to associate the preset field information with the preset text description information to obtain an association relationship, and generate page configuration information according to the association relationship.

[0186] In one example, the apparatus 50 further includes:

[0187] A processing module 509 is configured to segment the original text based on a preset large language model to obtain at least one text segment;

[0188] The first generating module 510 is configured to identify keywords in a text segment based on a preset large language model, and generate a preset knowledge base from the keywords and at least one text segment.

[0189] In one example, the apparatus 50 further includes:

[0190] The second generation module 511 is used to generate a preset index relationship library based on the preset large language model, if it is detected that there are other text addresses in the text segment, based on the association relationship between the text segment and the other text addresses in the text segment; wherein the other text addresses are used to represent the addresses of the texts linked to in the text segment.

[0191] The device provided in this embodiment can execute the method of any of the above embodiments, and its execution method and beneficial effects are similar, which will not be repeated here.

[0192] An embodiment of the present disclosure further provides a terminal device, which includes: a memory storing a computer program; a processor for executing the computer program, and when the computer program is executed by the processor, the method of any of the above embodiments can be implemented.

[0193] The present disclosure also provides a client, including:

[0194] A second acquisition module is configured to acquire local input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to query a preset knowledge base;

[0195] A second search module is configured to search a preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between text segments and other text segments;

[0196] The second determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate an output result.

[0197] The present disclosure also provides a server, including:

[0198] The third acquisition module is used to obtain input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; the first prompt word is used to query in a preset knowledge base; wherein the input information is sent by the client to the server;

[0199] a third search module configured to search a preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between text segments and other text segments;

[0200] The third determination module is used to generate a second prompt word based on the input information, the first text set and a preset prompt word template, input the second prompt word into the preset large language model, and determine the output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

[0201] For example, Figure 6 This is a schematic diagram of the structure of a terminal device in the embodiment of the present disclosure. Figure 6, which shows a schematic structural diagram of a terminal device 1000 suitable for implementing the embodiments of the present disclosure. The terminal device 1000 in the embodiments of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The terminal device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present disclosure.

[0202] like Figure 6 As shown, the terminal device 1000 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 1001, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1008 into a random access memory (RAM) 1003. Various programs and data required for the operation of the terminal device 1000 are also stored in the RAM 1003. The processing device 1001, the ROM 1002, and the RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0203] Typically, the following devices may be connected to the I / O interface 1005: an input device 1006 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 1007 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1008 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 may allow the terminal device 1000 to communicate with other devices wirelessly or by wire to exchange data. Figure 6 The terminal device 1000 is shown as having various devices, but it should be understood that it is not required to implement or possess all the devices shown, and more or fewer devices may be implemented or possessed instead.

[0204] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 1009, or installed from the storage device 1008, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.

[0205] It should be noted that the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, 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 above. In the present disclosure, a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component. In the present disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0206] In some embodiments, the client and server can communicate using any currently known or future developed network protocol, such as HTTP (HyperText Transfer Protocol), and can be interconnected with 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"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0207] The computer-readable medium may be included in the terminal device, or may exist independently without being incorporated into the terminal device.

[0208] The computer-readable medium carries one or more programs. When executed by the terminal device, the one or more programs cause the terminal device to: obtain input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access text; the text is used to represent content to be queried; the first prompt word is used to query a preset knowledge base; search the preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between text segments and other text segments; generate a second prompt word based on the input information, the first text set, and a preset prompt word template; input the second prompt word into a preset large language model; and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate an output result.

[0209] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including, but not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0210] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of the boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0211] The units involved in the embodiments described in this disclosure may be implemented in software or hardware, wherein the name of a unit does not necessarily limit the unit itself.

[0212] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, and without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chip (SOCs), complex programmable logic devices (CPLDs), and the like.

[0213] 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.

[0214] The embodiments of the present disclosure also provide a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the method of any of the above embodiments can be implemented. The execution method and beneficial effects are similar and will not be repeated here.

[0215] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device that includes the element.

[0216] The above are merely specific embodiments of the present disclosure, intended to enable those skilled in the art to understand and implement the present disclosure. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure is not limited to these embodiments, but is to be construed in the broadest manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for determining an output result, characterized in that: include: Obtaining input information, obtaining corresponding text from a preset text address, and generating a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent content to be queried; and the first prompt word is used to query the preset knowledge base; Searching a preset knowledge base for at least one text segment associated with the first prompt word, traversing a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generating a first text set based on the first traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between the text segment and other text segments; A second prompt word is generated based on the input information, the first text set, and a preset prompt word template, the second prompt word is input into a preset large language model, and an output result is determined based on the second prompt word and the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

2. The method according to claim 1, characterized in that The step of obtaining the corresponding text from the preset text address and generating the first prompt word according to the text and the input information includes: Obtaining corresponding text from a preset text address, and obtaining page configuration information corresponding to the text; wherein the page configuration information is pre-configured in the text; Based on the preset large language model, extracting data structure information and configuration rules of the text according to the page configuration information; The first prompt word is generated according to the input information, the data structure information and the configuration rule.

3. The method according to claim 2, characterized in that Generating the first prompt word according to the input information, the data structure information, and the configuration rule includes: Based on the preset large language model, extracting field information from the data structure information according to the input information; Searching the configuration rule for text description information associated with the field information, and determining the text description information as the first prompt word; wherein the configuration rule is used to characterize a mapping relationship between the field information and the text description information.

4. The method according to claim 1 or 3, characterized in that The step of traversing a preset index relationship library according to the at least one text segment to obtain a first traversal result, and generating a first text set according to the first traversal result, includes: Traversing each of the text segments in the preset index relationship library to obtain a first traversal result; If the first traversal result indicates that there are other text addresses associated with the text segment, the text segment and the other text addresses associated with the text segment are used to generate a target text segment; wherein the other text addresses associated with the text segment are used to represent addresses of other texts linked to the text segment; If the first traversal result indicates that there is no other text address associated with the text segment, generating a target text segment from the text segment; The first text set is generated by using at least one of the target text segments.

5. The method according to claim 4, characterized in that After generating the first text set from at least one target text segment, the method further includes: Calculating the similarity between the first prompt word and each target text segment in the first text set based on a preset similarity algorithm; The target text segment whose similarity satisfies a first preset condition is determined as the first target text.

6. The method according to claim 1, characterized in that The determining an output result according to the second prompt word based on the preset large language model includes: Outputting an initial result based on the preset large language model according to the second prompt word; wherein the initial result is used to represent a result initially obtained by the preset large language model according to the second prompt word; The output result is determined based on the preset large language model according to the matching degree between the initial result and the input information; wherein the output result is used to represent the target result determined based on the preset large language model.

7. The method according to claim 6, characterized in that The determining the output result based on the preset large language model according to the matching degree between the initial result and the input information includes: If the matching degree satisfies a second preset condition, the initial result is determined as the output result; wherein the second preset condition is that the matching degree is greater than a threshold; If the matching degree does not meet the second preset condition, other text addresses associated with the text fragment in the first text set are queried, and the output result is determined based on the other text addresses; wherein the other text addresses associated with the text fragment are used to represent the addresses of other texts linked to in the text fragment.

8. The method according to claim 7, characterized in that The determining the output result according to the other text address includes: Obtaining corresponding other text from the other text address, and generating a third prompt word based on the other text and the input information; wherein the other text address is used to access the other text; and the third prompt word is used to search in the preset knowledge base; Searching a preset knowledge base for at least one text segment associated with the third prompt word, traversing a preset index relationship library based on the at least one text segment to obtain a second traversal result, and generating a second text set based on the second traversal result; wherein the preset knowledge base is used to pre-store text segments; and the preset index relationship library is used to store associations between the text segment and other text segments; A fourth prompt word is generated based on the input information, the second text set, and a preset prompt word template, the fourth prompt word is input into a preset large language model, and an output result is determined based on the fourth prompt word based on the preset large language model; wherein the output result is used to be displayed on a terminal device.

9. The method according to claim 1, characterized in that The text segment is generated from a plurality of original texts.

10. The method according to claim 1, characterized in that The method further comprises: Extracting data structure information from the original text based on the preset large language model; Extracting preset field information from the data structure information based on the preset large language model; The preset field information and the preset text description information are associated to obtain an association relationship, and page configuration information is generated according to the association relationship.

11. The method according to claim 1, wherein The method further comprises: Segmenting the original text based on the preset large language model to obtain at least one text segment; Based on the preset large language model, keywords in the text segment are identified, and the keywords and the at least one text segment are combined to generate the preset knowledge base.

12. The method according to claim 11, characterized in that The method further comprises: Based on the preset large language model, if it is detected that other text addresses exist in the text fragment, the association relationship between the text fragment and the other text addresses in the text fragment is generated into the preset index relationship library; wherein the other text address is used to represent the address of the text linked to in the text fragment.

13. An output result determination device, characterized in that: include: A first acquisition module is configured to acquire input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to perform a query in the preset knowledge base; A first search module is configured to search a preset knowledge base for at least one text segment associated with the first prompt word, perform a traversal in a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between the text segment and other text segments; A first determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

14. A client, characterized in that: include: a second acquisition module, configured to acquire local input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; and the first prompt word is used to perform a query in the preset knowledge base; a second search module configured to search a preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between the text segment and other text segments; A second determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

15. A server, characterized in that: include: a third acquisition module, configured to acquire input information, obtain corresponding text from a preset text address, and generate a first prompt word based on the text and the input information; wherein the preset text address is used to access the text; the text is used to represent the content to be queried; the first prompt word is used to query the preset knowledge base; wherein the input information is sent by the client to the server; a third search module configured to search a preset knowledge base for at least one text segment associated with the first prompt word, traverse a preset index relationship library based on the at least one text segment to obtain a first traversal result, and generate a first text set based on the first traversal result; wherein the preset knowledge base is configured to pre-store text segments; and the preset index relationship library is configured to store associations between the text segment and other text segments; A third determination module is configured to generate a second prompt word based on the input information, the first text set, and a preset prompt word template, input the second prompt word into a preset large language model, and determine an output result based on the second prompt word based on the preset large language model; wherein the second prompt word is used to instruct the preset large language model to generate the output result.

16. A terminal device, characterized in that: include: A processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor performs the method according to any one of claims 1 to 12.

17. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 12 is implemented.

Images