Text evaluation method and device based on large model and electronic equipment
By using a large model-based text evaluation method, which generates evaluation-related responses using multiple prompt text templates and a large model, the inefficiency caused by the broad evaluation rules in existing technologies is solved, and more efficient and accurate text evaluation is achieved.
Patent Information
- Application Number
- CN202511006785.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-21
- Publication Date
- 2025-11-21
AI Technical Summary
Existing text evaluation methods have overly broad evaluation rules for cue texts, resulting in poor evaluation efficiency.
By identifying the text to be evaluated, the writing requirements text, and the evaluation-related questions, a large model is used to generate evaluation-related responses, thereby determining the evaluation results. This includes using multiple prompt text templates and a large model for text evaluation processing, improving the accuracy and efficiency of the evaluation.
It improves the accuracy and efficiency of text evaluation, ensures that the evaluation results meet writing requirements, and reduces errors from human experience-based judgment.
Smart Images
Figure CN120994773A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of artificial intelligence technology, particularly to the fields of deep learning, natural language processing, and large models, and especially to a text evaluation method, apparatus, and electronic device based on a large model. Background Technology
[0002] Current text evaluation methods require personnel to determine the cue texts used for evaluation based on experience; and then combine the cue texts with a large model to evaluate the text.
[0003] In the methods described above, the evaluation rules in the prompt text are rather broad, resulting in poor evaluation efficiency. Summary of the Invention
[0004] This disclosure provides a text evaluation method, apparatus, and electronic device based on a large model.
[0005] According to one aspect of this disclosure, a text evaluation method based on a large model is provided, the method comprising: determining a text to be evaluated, a writing requirement text corresponding to the text to be evaluated, and at least one evaluation-related question in the writing requirement text; determining at least one evaluation-related response corresponding to the at least one evaluation-related question based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and a first large model; and determining an evaluation result of the text to be evaluated based on the at least one evaluation-related response.
[0006] According to another aspect of this disclosure, a text evaluation apparatus based on a large model is provided. The apparatus includes: a first determining module, configured to determine a text to be evaluated, a writing requirement text corresponding to the text to be evaluated, and at least one evaluation-related question in the writing requirement text; a second determining module, configured to determine at least one evaluation-related response corresponding to the at least one evaluation-related question based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and a first large model; and a third determining module, configured to determine the evaluation result of the text to be evaluated based on the at least one evaluation-related response.
[0007] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to said at least one processor; wherein the memory stores instructions executable by said at least one processor, said instructions being executed by said at least one processor to enable said at least one processor to perform the large-model-based text evaluation method proposed above in this disclosure.
[0008] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions for causing a computer to execute the large-model-based text evaluation method proposed above in this disclosure.
[0009] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the large-model-based text evaluation method proposed above.
[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description
[0011] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:
[0012] Figure 1 This is a schematic diagram based on the first embodiment of the present disclosure;
[0013] Figure 2 This is a schematic diagram according to the second embodiment of the present disclosure;
[0014] Figure 3 This is a schematic diagram according to the third embodiment of the present disclosure;
[0015] Figure 4 This is a block diagram of an electronic device used to implement the large-model-based text evaluation method of the embodiments of this disclosure. Detailed Implementation
[0016] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.
[0017] Current text evaluation methods require personnel to determine the cue texts used for evaluation based on experience; and then combine the cue texts with a large model to evaluate the text.
[0018] In the methods described above, the evaluation rules in the prompt text are rather broad, resulting in poor evaluation efficiency.
[0019] To address the aforementioned issues, this disclosure proposes a text evaluation method, apparatus, and electronic device based on a large model.
[0020] Figure 1The diagram is based on the first embodiment of this disclosure. It should be noted that the text evaluation method based on a large model in this disclosure can be applied to a text evaluation device based on a large model. This device can be configured in an electronic device so that the electronic device can perform text evaluation functions.
[0021] Among them, electronic devices can be any device with computing capabilities, such as personal computers (PCs), mobile terminals, servers, etc. Mobile terminals can be, for example, in-vehicle devices, mobile phones, tablets, personal digital assistants, wearable devices, smart speakers, servers, server clusters, and other hardware devices with various operating systems, touch screens and / or displays.
[0022] The text evaluation device based on a large model can also be software in an electronic device, such as text evaluation software. The following embodiments use an electronic device as an example for illustration.
[0023] like Figure 1 As shown, this large-model-based text evaluation method may include the following steps:
[0024] Step 101: Identify the text to be evaluated, the corresponding writing requirements text, and at least one evaluation-related question from the writing requirements text.
[0025] In this embodiment of the disclosure, the text to be evaluated can be generated based on the writing requirement text. That is, the writing requirement text corresponding to the text to be evaluated can be the writing requirement text used when generating the text to be evaluated.
[0026] In this embodiment of the disclosure, the text to be evaluated can be text of various text types. For example, the text to be evaluated can be official document text, non-official document text, etc.
[0027] The document types can include at least one of the following: resolutions, decisions, orders, bulletins, announcements, notices, opinions, notifications, circulars, reports, requests for instructions, replies, proposals, letters, minutes, etc. There are no specific limitations here, and they can be set according to actual needs.
[0028] Different types of official documents have different requirements for text content and / or writing style. Therefore, different writing requirements can be set for different types of official documents.
[0029] Step 102: Based on the text to be evaluated, the writing requirements text, at least one evaluation-related question, and the first major model, determine at least one evaluation-related response corresponding to at least one evaluation-related question.
[0030] In this embodiment of the disclosure, the process of the electronic device performing step 102 may, for example, be: obtaining a second prompt text template; the second prompt text template is used for evaluating the generation process of relevant responses; determining a second prompt text based on the text to be evaluated, the writing requirement text, at least one evaluation-related question, and the second prompt text template; inputting the second prompt text into a first large model, and obtaining at least one evaluation-related response output by the first large model.
[0031] In this embodiment of the disclosure, the second prompt text template may include: a text description fragment indicating the generation of evaluation-related responses, a location for filling the text to be evaluated, a location for filling the writing requirement text, a location for filling at least one evaluation-related question, etc. The locations may be marked with text types so that the object or large model understands which types of text are used to fill that location.
[0032] The process by which the electronic device determines the second prompt text based on the text to be evaluated, the writing requirement text, at least one evaluation-related question, and the second prompt text template can be, for example, determining each position in the second prompt text template and filling in the corresponding text at each position to obtain the second prompt text.
[0033] Specifically, based on the text to be evaluated, the writing requirements text, at least one evaluation-related question, and the second prompt text template, the second prompt text is determined, and then the evaluation-related response generation process is carried out. This enables the first model to understand the generation requirements of the evaluation-related response, thereby generating the evaluation-related response and further improving the accuracy of the evaluation-related response.
[0034] In this embodiment of the disclosure, in order to facilitate understanding of the correspondence between at least one evaluation-related question and at least one evaluation-related response, the electronic device can output at least one evaluation-related question and the evaluation-related response corresponding to each evaluation-related question.
[0035] In this embodiment of the disclosure, to enhance the richness of the evaluation-related responses, more content is considered for determining the evaluation results of the text to be evaluated, thereby improving the accuracy of the evaluation results. The evaluation-related responses to the evaluation-related questions include at least one of the following: the answer to the evaluation-related question, the reasoning behind the answer, and suggestions for text improvement.
[0036] Among them, based on at least one of the reasons for the answer and the text improvement suggestions, the writing requirement text can be adjusted to improve the accuracy of the writing requirement text, and thus improve the accuracy of the text generated based on the writing requirement text.
[0037] Among them, based on at least one of the reasons for the response and suggestions for text improvement, the text to be evaluated can be adjusted to improve its accuracy.
[0038] In this embodiment of the disclosure, in order to further improve the richness of the evaluation results, the output of the first model may also include a thought chain, which is used to instruct the thought logic of the first model in the process of generating evaluation-related responses.
[0039] In this embodiment of the disclosure, to further improve the accuracy of the evaluation-related responses, the electronic device can guide the generation of evaluation-related responses by providing examples in the second prompt text template. Correspondingly, the second prompt text template includes examples of writing requirements text, examples of text to be evaluated, examples of evaluation-related questions, and examples of evaluation-related responses, used to guide the first model in generating at least one evaluation-related response.
[0040] The writing requirement states that the text type of the example text can be the same as or different from the text to be evaluated.
[0041] For example, the writing prompt might include the following text: "Analyze the following artists and their representative works: 1. Van Gogh – *Starry Night*; 2. Auguste Rodin – *The Thinker*. When an artist matches a work, determine whether they are a painter or a sculptor. If a painter, generate a description of their work, including the term 'Renaissance period.' If a sculptor, generate a brief introduction of the artist, approximately 50 words. When an artist does not match a work, determine whether they are a painter or a sculptor. If a painter, correct the incorrect work title, outputting 'The following is the correction.' If a sculptor, output 'Please re-enter.'"
[0042] For example, the text to be evaluated could be: "Van Gogh – *The Starry Night*. Van Gogh was a Dutch Post-Impressionist painter, whose representative works include *The Starry Night* and *Sunflowers*. *The Starry Night* is his most famous work, showcasing Van Gogh's unique style and use of color, and is one of the representative works of Post-Impressionist painting. Auguste Rodin – *The Thinker*. Auguste Rodin was a French sculptor, whose representative works include *The Thinker*. *The Thinker* is one of his representative works, employing a Cubist style, and expressing the spiritual connotation of human thought and exploration."
[0043] For example, evaluation examples of related questions could include: "Does the model-generated description of Van Gogh's works include the term 'Renaissance period'? Is the model-generated biographical text about Auguste Rodin around 50 words?"
[0044] For example, evaluating relevant response examples could be: {{'query':"Does the model-generated description of Van Gogh's works include the term 'Renaissance period'?",'score':0,'reason':"The description is missing the term 'Renaissance period'"}}, {{"query":"Is the model-generated biographical text about Auguste Rodin around 50 words?",'score':0,'reason':"Auguste Rodin's biographical text exceeds 60 words"}}.
[0045] Step 103: Determine the evaluation result of the text to be evaluated based on at least one evaluation-related response.
[0046] In this embodiment of the disclosure, the process of the electronic device performing step 103 may, for example, be: determining an evaluation score for at least one evaluation-related question based on at least one evaluation-related response; determining a total evaluation score for the text to be evaluated based on at least one evaluation score; and determining an evaluation result based on at least one of the at least one evaluation-related response, at least one evaluation score, and the total evaluation score.
[0047] In one example of this disclosure, the process by which an electronic device determines at least one evaluation score may be, for example, inputting at least one evaluation-related question, the evaluation-related answer corresponding to the evaluation-related question, and the writing requirement text into a large model for score evaluation, and obtaining at least one evaluation score output by the large model.
[0048] In another example, the assessment-related question could be a judgment question, and the electronic device could determine at least one assessment score based on the judgment result in at least one assessment-related response. For example, if the judgment result is yes, the assessment score is a first value; if the judgment result is no, the assessment score is a second value.
[0049] In this process, determining the assessment score for at least one assessment-related question based on at least one assessment-related response, and then determining the total assessment score, thereby obtaining the assessment result, can improve the accuracy of the obtained assessment result.
[0050] The text evaluation method based on a large model in this disclosure involves determining the text to be evaluated, the corresponding writing requirement text, and at least one evaluation-related question from the writing requirement text; determining at least one evaluation-related response corresponding to the at least one evaluation-related question based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and a first large model; and determining the evaluation result of the text to be evaluated based on the at least one evaluation-related response. The at least one evaluation-related question accurately reflects the evaluation rules, thereby improving the accuracy of the evaluation result and further enhancing evaluation efficiency.
[0051] To further ensure that the assessment-related questions reflect the assessment rules and improve the matching degree between the assessment-related questions and the text to be assessed, thereby further improving the efficiency of text assessment, the writing requirement text corresponding to the text to be assessed can be determined; based on the writing requirement text and the second major model, at least one assessment-related question can be determined. For example... Figure 2 As shown, Figure 2 This is a schematic diagram based on the second embodiment of the present disclosure. Figure 2 The illustrated embodiment may include the following steps:
[0052] Step 201: Determine the text to be evaluated and the corresponding writing requirement text; wherein, the text to be evaluated is generated based on the writing requirement text.
[0053] The text to be evaluated can be generated based on the writing requirements text. In other words, the writing requirements text corresponding to the text to be evaluated can be the writing requirements text used when generating the text to be evaluated.
[0054] The text to be evaluated can be any type of text. For example, the text to be evaluated can be official documents or non-official documents.
[0055] In this embodiment of the disclosure, before step 202, or after step 201, to further improve the accuracy of the writing requirement text, error types can be described in the writing requirement text to avoid such error types during text generation; or, the error types can be considered during text evaluation. The electronic device can also perform the following processes: determining the text type to which the text to be evaluated belongs; determining the error types in multiple sample texts under the text type; and adjusting the writing requirement text according to the error types to obtain the adjusted writing requirement text.
[0056] In this embodiment of the disclosure, the process of adjusting the writing requirement text according to the error type to obtain the adjusted writing requirement text by the electronic device may be as follows: obtaining a third prompt text template; the third prompt text template is used for adjusting the writing requirement text; determining the third prompt text according to the writing requirement text, the error type and the third prompt text template; inputting the third prompt text into the third model, and obtaining the adjusted writing requirement text output by the third model.
[0057] In particular, by combining the third major model and error types to adjust the writing requirements text, the accuracy of the adjusted writing requirements text can be further improved.
[0058] Step 202: Based on the writing requirements text and the second major model, identify at least one assessment-related question.
[0059] In this embodiment of the disclosure, to further improve the accuracy of the evaluation-related questions output by the second model, the prompt text template and the writing requirement text can be combined to determine the prompt text. This allows the second model to understand the generation requirements of the evaluation-related questions and thus generate them. Correspondingly, the process by which the electronic device determines at least one evaluation-related question based on the writing requirement text and the second model can, for example, involve: obtaining a first prompt text template; using the first prompt text template for the generation of evaluation-related questions; determining the first prompt text based on the writing requirement text and the first prompt text template; inputting the first prompt text into the second model to obtain at least one evaluation-related question output by the second model.
[0060] In this embodiment, the first prompt text template may include: a text description fragment instructing the generation of evaluation-related questions, a location for filling the writing requirement text, etc. The location for filling the writing requirement text may be marked so that the object or large model understands that the location is for filling the writing requirement text. Correspondingly, the process by which the electronic device determines the first prompt text based on the writing requirement text and the first prompt text template may, for example, involve determining the location in the first prompt text template for filling the writing requirement text; and filling the writing requirement text at that location to obtain the first prompt text.
[0061] In this embodiment of the disclosure, when the text to be evaluated is an official document, at least one evaluation-related question can be an evaluation-related question related to the evaluation rules of the official document. For example, at least one evaluation-related question includes: at least one content-related question on the text content dimension, and / or, at least one normative question on the writing style dimension.
[0062] The content-related questions in the text content dimension are used to inquire or determine whether the text to be evaluated contains certain content required for official documents. Examples of content-related questions in the text content dimension include at least one of the following: "Does it fully cover the five core parts required for a request-type official document (necessity / maturity conditions / overall approach / establishment steps / requested matters)?"; "Does the necessity justification include the three dimensions of mitigating financial risks, improving capital adequacy ratio, and brand reshaping?"; "Does the maturity conditions section fully present the three elements of policy basis, economic foundation, and operational feasibility?"; "Does the overall approach clearly include core elements such as guiding ideology, objectives, and establishment methods?"; "Does the establishment steps present a complete chain of asset verification → asset disposal → capital increase and share expansion → approval → listing → investment attraction → IPO?"
[0063] Among these, the normative questions in the document writing standardization dimension are used to inquire or determine whether the text to be evaluated conforms to certain normative writing standards required for official documents. These normative questions include at least one of the following: "Does the title conform to the standard format of 'issuing authority + subject matter + document type'?", "Does the body structure strictly follow the three-part logic of 'reason for request + matter requested + conclusion'?", "Does it appropriately use request-specific closing phrases such as 'Please approve or disapprove'?", and "Is the use of professional terminology (such as 'asset verification' and 'capital increase') accurate and standardized?"
[0064] The inclusion of at least one content-related question in the text content dimension and / or at least one norm-related question in the document writing norms dimension can further improve the matching degree between the evaluation-related questions and the document text to be evaluated, thereby further improving the evaluation efficiency of the document text.
[0065] Step 203: Based on the text to be evaluated, the writing requirements text, at least one evaluation-related question, and the first major model, determine at least one evaluation-related response corresponding to at least one evaluation-related question.
[0066] Step 204: Determine the evaluation result of the text to be evaluated based on at least one evaluation-related response.
[0067] It should be noted that the various major models involved in this disclosure, such as the first major model, the second major model, the third major model, etc., can be the same major model or different major models. No specific limitation is made here, and they can be set according to actual needs.
[0068] It should be noted that for details of steps 203 to 204, please refer to [the relevant documentation / reference]. Figure 1 Steps 102 to 103 in the illustrated embodiment will not be described in detail here.
[0069] The text evaluation method based on a large model disclosed in this embodiment determines the text to be evaluated and the corresponding writing requirement text; wherein the text to be evaluated is generated based on the writing requirement text; based on the writing requirement text and a second large model, at least one evaluation-related question is determined; based on the text to be evaluated, the writing requirement text, at least one evaluation-related question, and a first large model, at least one evaluation-related response corresponding to the at least one evaluation-related question is determined; based on the at least one evaluation-related response, the evaluation result of the text to be evaluated is determined; wherein the writing requirement text corresponding to the text to be evaluated is determined; based on the writing requirement text and the second large model, at least one evaluation-related question is determined, which can further ensure that the evaluation-related questions can reflect the evaluation rules, improve the matching degree between the evaluation-related questions and the text to be evaluated, and thus further improve the efficiency of text evaluation.
[0070] To implement the above embodiments, this disclosure also provides a text evaluation device based on a large model. For example... Figure 3 As shown, Figure 3 This is a schematic diagram according to a third embodiment of the present disclosure. The large-model-based text evaluation device 30 may include: a first determining module 301, a second determining module 302, and a third determining module 303.
[0071] The first determining module 301 is used to determine the text to be evaluated, the writing requirement text corresponding to the text to be evaluated, and at least one evaluation-related question in the writing requirement text; the second determining module 302 is used to determine at least one evaluation-related response corresponding to the at least one evaluation-related question based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and the first large model; the third determining module 303 is used to determine the evaluation result of the text to be evaluated based on the at least one evaluation-related response.
[0072] As one possible implementation of this disclosure, the first determining module 301 includes a first determining unit and a second determining unit; the first determining unit is used to determine the text to be evaluated and the writing requirement text corresponding to the text to be evaluated; wherein the text to be evaluated is generated based on the writing requirement text; the second determining unit is used to determine the at least one evaluation-related question according to the writing requirement text and the second model.
[0073] As one possible implementation of this disclosure, the second determining unit is specifically used to: obtain a first prompt text template; the first prompt text template is used for the generation and processing of evaluation-related questions; determine a first prompt text based on the writing requirement text and the first prompt text template; input the first prompt text into the second large model, and obtain the at least one evaluation-related question output by the second large model.
[0074] As one possible implementation of this disclosure, the second determining module 302 is specifically used to: obtain a second prompt text template; the second prompt text template is used for evaluating the generation process of relevant responses; determine a second prompt text based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and the second prompt text template; input the second prompt text into the first large model, and obtain the at least one evaluation-related response output by the first large model.
[0075] As one possible implementation of this disclosure, the second prompt text template includes a writing requirement text example, a text example to be evaluated, an evaluation-related question example, and an evaluation-related response example, which are used to guide the first large model to generate at least one evaluation-related response.
[0076] As one possible implementation of this disclosure, the evaluation-related responses to the evaluation-related questions include at least one of the following: the answer to the evaluation-related question, the reasoning behind the answer, and suggestions for text improvement.
[0077] As one possible implementation of this disclosure, the apparatus further includes: a fourth determining module, a fifth determining module, and an adjustment processing module; the fourth determining module is used to determine the text type to which the text to be evaluated belongs; the fifth determining module is used to determine the error type among multiple sample texts under the text type; the adjustment processing module is used to adjust the writing requirement text according to the error type to obtain the adjusted writing requirement text.
[0078] As one possible implementation of this disclosure, the adjustment processing module is specifically used to: obtain a third prompt text template; the third prompt text template is used for adjusting the writing requirement text; determine a third prompt text based on the writing requirement text, the error type, and the third prompt text template; input the third prompt text into a third model, and obtain the adjusted writing requirement text output by the third model.
[0079] As one possible implementation of this disclosure, the third determining module 303 is specifically configured to: determine the evaluation score of the at least one evaluation-related question based on the at least one evaluation-related response; determine the total evaluation score of the text to be evaluated based on the at least one evaluation score; and determine the evaluation result based on at least one of the at least one evaluation-related response, the at least one evaluation score, and the total evaluation score.
[0080] As one possible implementation of this disclosure, the text to be evaluated includes official document text; the at least one evaluation-related issue includes at least one content-related issue in the dimension of text content, and / or at least one norm-related issue in the dimension of writing style.
[0081] The large-model-based text evaluation device of this disclosure determines the text to be evaluated, the corresponding writing requirement text, and at least one evaluation-related question in the writing requirement text; determines at least one evaluation-related response corresponding to the at least one evaluation-related question based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and a first large model; and determines the evaluation result of the text to be evaluated based on the at least one evaluation-related response. The at least one evaluation-related question accurately reflects the evaluation rules, thereby improving the accuracy of the evaluation result and further enhancing evaluation efficiency.
[0082] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information are all carried out with the consent of the users, and all comply with the provisions of relevant laws and regulations, and do not violate public order and good morals.
[0083] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.
[0084] Figure 4 A schematic block diagram of an example electronic device 400 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.
[0085] like Figure 4 As shown, device 400 includes a computing unit 401, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 402 or a computer program loaded from storage unit 408 into random access memory (RAM) 403. RAM 403 may also store various programs and data required for the operation of device 400. The computing unit 401, ROM 402, and RAM 403 are interconnected via bus 404. Input / output (I / O) interface 405 is also connected to bus 404.
[0086] Multiple components in device 400 are connected to I / O interface 405, including: input unit 406, such as keyboard, mouse, etc.; output unit 407, such as various types of monitors, speakers, etc.; storage unit 408, such as disk, optical disk, etc.; and communication unit 409, such as network card, modem, wireless transceiver, etc. Communication unit 409 allows device 400 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0087] The computing unit 401 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above, such as large model-based text evaluation methods. For example, in some embodiments, the large model-based text evaluation method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed on device 400 via ROM 402 and / or communication unit 409. When the computer program is loaded into RAM 403 and executed by the computing unit 401, one or more steps of the large model-based text evaluation method described above can be performed. Alternatively, in other embodiments, computing unit 401 may be configured to perform a large model-based text evaluation method by any other suitable means (e.g., by means of firmware).
[0088] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0089] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0090] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0091] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0092] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.
[0093] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.
[0094] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.
[0095] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.
Claims
1. A text evaluation method based on a large model, the method comprising: Identify the text to be evaluated, the corresponding writing requirements text, and at least one evaluation-related question from the writing requirements text; Based on the text to be evaluated, the writing requirements text, the at least one evaluation-related question, and the first major model, determine at least one evaluation-related response corresponding to the at least one evaluation-related question; The evaluation result of the text to be evaluated is determined based on the at least one evaluation-related response.
2. The method according to claim 1, wherein, The process of determining the text to be evaluated, the corresponding writing requirement text, and at least one evaluation-related question from the writing requirement text includes: Determine the text to be evaluated and the corresponding writing requirement text; wherein the text to be evaluated is generated based on the writing requirement text; Based on the writing requirements text and the second major model, determine at least one evaluation-related question.
3. The method according to claim 2, wherein, The step of determining at least one evaluation-related question based on the writing requirements text and the second major model includes: Obtain the first prompt text template; the first prompt text template is used to evaluate the generation and processing of related questions. Based on the writing requirements text and the first prompt text template, determine the first prompt text; Input the first prompt text into the second large model, and obtain the at least one evaluation-related question output by the second large model.
4. The method according to claim 1, wherein, The step of determining at least one assessment-related response corresponding to the at least one assessment-related question based on the text to be assessed, the writing requirement text, the at least one assessment-related question, and the first major model includes: Obtain the second prompt text template; the second prompt text template is used to evaluate the generation and processing of relevant responses; The second prompt text is determined based on the text to be evaluated, the writing requirement text, the at least one evaluation-related question, and the second prompt text template; Input the second prompt text into the first large model and obtain the at least one evaluation-related response output by the first large model.
5. The method according to claim 4, wherein, The second prompt text template includes examples of writing requirements text, examples of text to be evaluated, examples of evaluation-related questions, and examples of evaluation-related responses, which are used to guide the first large model to generate at least one evaluation-related response.
6. The method according to claim 1 or 4, wherein, The evaluation-related responses to the evaluation-related questions include at least one of the following: the answer to the evaluation-related question, the reasoning behind the answer, and suggestions for text improvement.
7. The method according to claim 2, wherein, The method further includes: Determine the text type to which the text to be evaluated belongs; Determine the error type in multiple sample texts under the given text type; The writing requirement text is adjusted according to the error type to obtain the adjusted writing requirement text.
8. The method according to claim 7, wherein, The step of adjusting the writing requirement text according to the error type to obtain the adjusted writing requirement text includes: Obtain the third prompt text template; the third prompt text template is used for adjusting and processing the writing requirement text; Based on the writing requirements text, the error type, and the third prompt text template, determine the third prompt text; Input the third prompt text into the third major model to obtain the adjusted writing requirement text output by the third major model.
9. The method according to claim 1, wherein, The step of determining the evaluation result of the text to be evaluated based on the at least one evaluation-related response includes: Based on the at least one assessment-related response, determine the assessment score for the at least one assessment-related question; Based on the at least one evaluation score, determine the total evaluation score of the text to be evaluated; The evaluation result is determined based on at least one of the at least one evaluation-related response, the at least one evaluation score, and the total evaluation score.
10. The method according to claim 1, wherein, The text to be evaluated includes official document texts; The at least one evaluation-related issue includes: at least one content-related issue in the text content dimension, and / or at least one norm-related issue in the writing style dimension.
11. A text evaluation device based on a large model, the device comprising: The first determining module is used to determine the text to be evaluated, the writing requirement text corresponding to the text to be evaluated, and at least one evaluation-related question in the writing requirement text; The second determining module is used to determine at least one assessment-related response corresponding to the at least one assessment-related question based on the text to be assessed, the writing requirement text, the at least one assessment-related question, and the first major model; The third determining module is used to determine the evaluation result of the text to be evaluated based on the at least one evaluation-related response.
12. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 10.
13. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1 to 10.
14. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 10.