Game role setting quantitative evaluation method and system based on large language model

Abstract

The invention provides a game role setting quantitative evaluation method and system based on a large language model, and the method comprises the steps: injecting a rule of a quantitative evaluation framework into a large language model through a preset cue word template, and inputting a game role setting CSV file in the large language model; performing comprehensive integrity evaluation on game role setting based on a double-layer integrity evaluation system, and performing quantitative analysis on character characteristics of roles through three-dimensional coordinates; based on the five-dimensional evaluation model, all-directional quantitative evaluation is carried out on game role setting to obtain a five-dimensional radar map, different weight coefficients are given to role behavior evidences with different credibility, and quantitative analysis is carried out; performing NPC visual angle evaluation, player visual angle evaluation and designer visual angle evaluation on game role setting, and calculating cognitive reversal degrees of different visual angles; and outputting an evaluation report by the large language model. According to the method, standardization, quantifiability and reproducibility of game role evaluation are achieved, and the requirement of a professional game development team for game role quality control is met.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence and game development technology, specifically to a quantitative evaluation method and system for game character settings based on a large language model. Background Technology

[0002] In the current gaming industry, character design is one of the core aspects of game development. The design of game characters directly affects player experience and game commercial performance. Traditional methods for evaluating character design have the following problems: First, traditional character setting mostly relies on subjective experience and judgment, lacking objective quantitative standards, and different evaluators have different evaluation results for the same character; Secondly, the traditional role setting evaluation process is not reproducible, the evaluation basis cannot be traced, and it is difficult to form an accumulative evaluation knowledge base; Furthermore, traditional character design cannot systematically evaluate the multi-dimensional characteristics of a character (psychological, appearance, social, functional, narrative), resulting in a single evaluation dimension; Traditional character design lacks quantitative analysis of the differences in perception from different perspectives (NPC first encounter, player experience, designer intent), and the completeness of traditional character design materials lacks clear evaluation standards, making it difficult to accurately guide subsequent supplementary work.

[0003] While existing technologies (such as traditional design document review processes and expert review meetings) can provide qualitative opinions, they cannot solve the aforementioned problems of quantitative evaluation, standardized processes, and reproducibility.

[0004] Based on the problems existing in the current technology, this application proposes a quantitative evaluation method and system for game character settings based on a large language model. Summary of the Invention

[0005] This application addresses one or more technical deficiencies in the prior art by proposing the following technical solution.

[0006] Based on the first aspect of this application, a quantitative evaluation method for game character design based on a large language model is proposed, including: S1: Inject the rules of the quantitative evaluation framework into the large language model through the preset prompt word template, and input the game character setting CSV file into the large language model; S2: The large language model is based on a two-layer completeness assessment system to comprehensively assess the completeness of game character settings, and uses three-dimensional coordinates to quantitatively analyze character personality traits; S3: The large language model is based on a five-dimensional evaluation model to conduct a comprehensive quantitative evaluation of the game character settings, resulting in a five-dimensional radar chart, and assigning different weight coefficients to character behavior evidence of different credibility and conducting quantitative analysis. S4: The large language model evaluates the game character settings from NPC perspective, player perspective, and designer perspective, and calculates the cognitive reversal degree from different perspectives. The calculation formula is as follows: ; Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings. S5: The large language model performs automated reasoning and calculation, and outputs an evaluation report including comprehensive completeness, quantitative evaluation analysis results, cognitive reversal degree and five-dimensional radar chart.

[0007] This application enables comprehensive and automated analysis and evaluation of game character settings, significantly improving the efficiency and objectivity of quantitative evaluation and providing game development teams with accurate and intuitive decision support.

[0008] Furthermore, the two-layer integrity assessment system includes a hard integrity assessment based on objective indicators and a perspective integrity assessment based on the perspectives of different cognitive subjects. The objective metrics include total text volume, number of dialogues, and number of scenes; The perspective integrity assessment includes evaluating the appearance, identity, initial behavior, and surface personality of NPCs; evaluating the plot performance, in-depth dialogue, behavioral changes, and interactive feedback from the player's perspective; and evaluating the background story, intrinsic motivation, design intentions, and deep settings from the designer's perspective.

[0009] This step transforms the subjective assessment of material sufficiency into a structured quantitative evaluation, which can accurately pinpoint the deficiencies in the game character design documents in terms of content and perspective, reducing the ambiguity and arbitrariness of the evaluation process.

[0010] Furthermore, the formula for calculating the overall completeness assessment is as follows: S0=W1 S1+W2 S2+W3 S3+W4 S4; Wherein, S0 represents the overall completeness score, S1 represents the hard completeness score, S2 represents the NPC perspective score, S3 represents the player perspective score, S4 represents the designer perspective score, and W1, W2, W3 and W4 represent the weights of the hard completeness score, the NPC perspective score, the player perspective score, and the designer perspective score, respectively.

[0011] Furthermore, the three-dimensional coordinates include a first coordinate axis representing the character's way of thinking, a second coordinate axis representing the character's emotional temperature, and a third coordinate axis representing the intensity of the character's performance. Each coordinate axis is divided into several numerical intervals. Combining the numerical intervals of each coordinate axis forms multiple personality positioning types. This step enables precise characterization and visual positioning of game characters' inner personalities, facilitating the classification, positioning, and comparative analysis of game characters.

[0012] Furthermore, the five-dimensional evaluation model includes psychological, appearance, social, functional, and narrative dimensions, assigning different weights to different dimensions and evaluating the quality of game characters. The five-dimensional evaluation model assesses the depth and complexity of game character personality development through the psychological dimension, the recognizability and completeness of game character visual design through the appearance dimension, the richness and diversity of game character relationship design through the social dimension, the clarity and value of game function positioning through the functional dimension, and the depth and tension of story potential through the narrative dimension.

[0013] This step can intuitively explain the advantages and disadvantages of game character design, enabling designers to optimize game characters in a targeted manner and conduct multi-character balance analysis.

[0014] Furthermore, the role behavior evidence includes role behavior evidence under extreme situations, role behavior evidence under stress, role behavior evidence of daily performance, and role behavior evidence of surface information.

[0015] Furthermore, evidence of character behavior in extreme situations refers to the behavior of game characters in life-or-death situations and core conflict scenarios; evidence of character behavior under stress refers to the stress state of game characters in situations of being alone in reality, fatigue reactions, and emotional outbursts; evidence of character behavior in daily performance refers to the daily performance of game characters under repetitive habits, stable patterns, and normal behaviors; and evidence of character behavior in surface information refers to the surface information of characters under polite behavior, professional behavior, and social masks.

[0016] This step distinguishes the differences in the contribution of behavioral evidence to revealing a character's true personality in different contexts, making the evaluation results closer to human cognitive patterns. The evaluation process is highly transparent and traceable, improving the interpretability of the evaluation results.

[0017] Based on the second aspect of this application, a quantitative evaluation system for game character design based on a large language model is also proposed, including: Preset module: The rules of the quantitative evaluation framework are injected into the large language model through preset prompt word templates, and the game character setting CSV file is input into the large language model; First analysis module: The large language model uses a two-layer completeness assessment system to comprehensively evaluate the completeness of game character settings and uses three-dimensional coordinates to quantitatively analyze character personality traits; The second analysis module: The large language model performs a comprehensive quantitative evaluation of the game character settings based on the five-dimensional evaluation model, obtains a five-dimensional radar chart, and assigns different weight coefficients to character behavior evidence with different credibility and performs quantitative analysis. The third analysis module: The large language model evaluates the game character settings from NPC perspective, player perspective, and designer perspective, and calculates the cognitive reversal degree from different perspectives. The calculation formula is as follows: ; Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings. Output module: The large language model performs automated reasoning and calculation, and outputs an evaluation report including comprehensive completeness, quantitative evaluation analysis, cognitive reversal degree and five-dimensional radar chart.

[0018] Furthermore, the system executes specific tasks according to a variety of preset workflow modes, including: a complete evaluation workflow for fully evaluating existing roles, an auxiliary filling workflow to guide the creation of new roles, a diagnostic optimization workflow to identify logical problems and provide improvement suggestions, a format conversion workflow to convert unstructured source documents into standardized formats, and a source document review workflow to assess the quality of the source documents before conversion and generate a rating report.

[0019] Based on a third aspect of this application, a computer program product is also proposed, having one or more computer programs thereon that, when executed by a computer processor, implement the method described above.

[0020] The technical advantages of this application are as follows: This application transforms subjective character setting evaluation into measurable values ​​through objective quantitative standards, utilizes cognitive differences from different perspectives for quantitative analysis, systematically evaluates the multi-dimensional characteristics of characters, and the evaluation results are objective and comparable. It achieves standardization, quantification, and reproducibility of game character setting evaluation, meets the needs of professional game development teams for quality control of game characters, can accurately locate the types and priorities of missing materials, reduces the rate of blind material supplementation, and forms a reusable evaluation knowledge base through source document review and real-time log recording. It can support the accumulation of team experience and provide a complete chain of evidence for source review and quality verification of each score. Attached Figure Description

[0021] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings.

[0022] Figure 1 This is a flowchart of a quantitative evaluation method for game character settings based on a large language model, provided according to an embodiment of this application.

[0023] Figure 2 This is a block diagram of a quantitative evaluation system for game character settings based on a large language model, provided according to an embodiment of this application.

[0024] Figure 3 This is a schematic diagram of the structure of a computer system suitable for implementing the electronic devices of the present application embodiments. Detailed Implementation

[0025] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0026] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0027] Figure 1 This application presents a quantitative evaluation method for game character settings based on a large language model, including: S1: Inject the rules of the quantitative evaluation framework into the large language model through the preset prompt word template, and input the game character setting CSV file into the large language model; S2: The large language model is based on a two-layer completeness assessment system to comprehensively assess the completeness of game character settings, and uses three-dimensional coordinates to quantitatively analyze character personality traits; S3: The large language model is based on a five-dimensional evaluation model to conduct a comprehensive quantitative evaluation of the game character settings, resulting in a five-dimensional radar chart, and assigning different weight coefficients to character behavior evidence of different credibility and conducting quantitative analysis. S4: The large language model evaluates the game character settings from NPC perspective, player perspective, and designer perspective, and calculates the cognitive reversal degree from different perspectives. The calculation formula is as follows: ; Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings. S5: The large language model performs automated reasoning and calculation, and outputs an evaluation report including comprehensive completeness, quantitative evaluation analysis results, cognitive reversal degree and five-dimensional radar chart.

[0028] It should be noted that this application enables comprehensive and automated analysis and evaluation of game character settings, significantly improving the efficiency and objectivity of the evaluation and providing accurate and intuitive decision support for game development teams.

[0029] It should be noted that the two-layer integrity assessment system includes a hard integrity assessment based on objective indicators and a perspective integrity assessment based on the perspectives of different cognitive subjects. The objective metrics include total text volume, number of dialogues, and number of scenes; The perspective integrity assessment includes evaluating the appearance, identity, initial behavior, and surface personality of NPCs; evaluating the plot performance, in-depth dialogue, behavioral changes, and interactive feedback from the player's perspective; and evaluating the background story, intrinsic motivation, design intentions, and deep settings from the designer's perspective.

[0030] It should be noted that this step transforms the subjective assessment of material sufficiency into a structured evaluation, which can accurately pinpoint the deficiencies in the game character setting documents in terms of content and perspective, reducing the ambiguity and arbitrariness of the evaluation process.

[0031] It should be noted that the calculation formula for the comprehensive integrity assessment is as follows: S0=W1 S1+W2 S2+W3 S3+W4 S4; Wherein, S0 represents the overall completeness score, S1 represents the hard completeness score, S2 represents the NPC perspective score, S3 represents the player perspective score, S4 represents the designer perspective score, and W1, W2, W3 and W4 represent the weights of the hard completeness score, the NPC perspective score, the player perspective score, and the designer perspective score, respectively.

[0032] It should be noted that the three-dimensional coordinates include a first coordinate axis representing the character's way of thinking, a second coordinate axis representing the character's emotional temperature, and a third coordinate axis representing the intensity of the character's performance. Each coordinate axis is divided into several numerical intervals. Combining the numerical intervals of each coordinate axis forms multiple personality positioning types. It should be noted that this step enables the precise portrayal and visual positioning of the game character's inner personality, facilitating the classification, positioning, and comparative analysis of the game character.

[0033] It should be noted that the five-dimensional evaluation model includes psychological, appearance, social, functional, and narrative dimensions, assigning different weights to different dimensions to evaluate the quality of game characters; The five-dimensional evaluation model assesses the depth and complexity of game character personality development through the psychological dimension, the recognizability and completeness of game character visual design through the appearance dimension, the richness and diversity of game character relationship design through the social dimension, the clarity and value of game function positioning through the functional dimension, and the depth and tension of story potential through the narrative dimension.

[0034] It should be noted that this step can intuitively explain the advantages and disadvantages of game character design, allowing designers to make targeted optimizations to game characters and conduct multi-character balance analysis.

[0035] It should be noted that the role behavior evidence includes role behavior evidence under extreme situations, role behavior evidence under stress, role behavior evidence in daily performance, and role behavior evidence based on surface information.

[0036] It should be noted that evidence of character behavior in extreme situations refers to the behavior of game characters in life-or-death situations and core conflict scenarios; evidence of character behavior under stress refers to the stress state of game characters in situations of being alone in reality, fatigue reactions, and emotional outbursts; evidence of character behavior in daily performance refers to the daily performance of game characters under repetitive habits, stable patterns, and normal behaviors; and evidence of character behavior in surface information refers to the surface information of characters under polite behavior, professional behavior, and social masks.

[0037] It should be noted that this step distinguishes the differences in the contribution of behavioral evidence to revealing the true personality of a character in different contexts, making the evaluation results closer to the laws of human cognition. The evaluation process is highly transparent and traceable, improving the interpretability of the evaluation results.

[0038] It should be noted that this application transforms subjective character setting evaluations into measurable values ​​through objective quantitative standards. It utilizes cognitive differences from different perspectives for quantitative analysis, systematically assesses the multi-dimensional characteristics of characters, and the evaluation results are objective and comparable. This achieves standardization, quantification, and reproducibility of game character setting evaluation, meeting the needs of professional game development teams for quality control of game characters. It can accurately locate the types and priorities of missing materials, reduce the rate of blindly supplementing materials, and form a reusable evaluation knowledge base through source document review and real-time log recording. This supports the accumulation of team experience, and each score has a complete chain of evidence for traceability review and quality verification.

[0039] The following is for reference. Figure 2It demonstrates a quantitative evaluation system for game character settings based on a large language model, including a preset module a, a first analysis module b, a second analysis module c, a third analysis module d, and an output module e.

[0040] In a specific embodiment, the preset module a is configured to: inject the rules of the quantitative evaluation framework into the large language model through a preset prompt word template, and input the game character setting CSV file into the large language model.

[0041] In a specific embodiment, the first analysis module b is configured to: the large language model performs a comprehensive integrity assessment of the game character settings based on a two-layer integrity assessment system, and performs quantitative analysis of the character's personality traits through three-dimensional coordinates.

[0042] In a specific embodiment, the second analysis module c is configured to: the large language model performs a comprehensive quantitative evaluation of the game character settings based on a five-dimensional evaluation model to obtain a five-dimensional radar chart, and assign different weight coefficients to character behavior evidence with different credibility and perform quantitative analysis.

[0043] In a specific embodiment, the third analysis module d is configured to: perform NPC perspective evaluation, player perspective evaluation, and designer perspective evaluation on the game character settings using the large language model, and calculate the cognitive reversal degree of different perspectives, using the following formula: ; Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings.

[0044] In a specific embodiment, the output module e is configured to: perform automated reasoning and calculation on the large language model, and output an evaluation report including comprehensive completeness, quantitative evaluation analysis, cognitive reversal degree and five-dimensional radar chart.

[0045] In a specific embodiment, the two-tiered integrity assessment system comprises two levels: hard integrity (objective standards) and perspective integrity (targeted assessment). Based on three objective indicators—total text (≥300 characters), number of dialogues (≥5 sentences), and number of scenes (≥3 scenes)—a weighted score is calculated to determine the hard completeness score. The evaluation process includes assessing the sufficiency of materials from the NPC's perspective, including appearance, identity, initial behavior, and surface personality; assessing the sufficiency of materials from the player's perspective, including plot presentation, in-depth dialogue, behavioral changes, and interactive feedback; and assessing the sufficiency of materials from the designer's perspective, including background story, intrinsic motivation, design intention, and deep settings, resulting in scores for the NPC's perspective, player's perspective, and designer's perspective, respectively. Calculate the overall completeness score: S0 = 0.4 S1+0.2 S2+0.25 S3+0.15 S4; S0 represents the overall completeness score, S1 represents the hard completeness score, S2 represents the NPC perspective score, S3 represents the player perspective score, and S4 represents the designer perspective score.

[0046] In a specific embodiment, the three-dimensional coordinates include a first coordinate axis representing the character's way of thinking, a second coordinate axis representing the character's emotional temperature, and a third coordinate axis representing the intensity of the character's performance. The first coordinate axis, the X-axis, is used to evaluate the basis for the game character's decision-making, including logical reasoning or emotional association; The second coordinate axis, the Y-axis, is used to assess the game character's willingness and behavior in actively establishing interpersonal connections; The third coordinate axis, Z-axis, is used to assess the degree to which the emotions and personality traits of game characters are manifested. Each coordinate axis (0-100) is divided into three intervals: high interval (67-100), medium interval (34-66), and low interval (0-33). Combining the three intervals of each coordinate axis forms 27 personality positioning types. In a specific embodiment, the five-dimensional evaluation model includes psychological, appearance, social, functional, and narrative dimensions. Different weights are assigned to different dimensions to evaluate the quality of game characters. For example, the psychological dimension is assigned a 40% weight to evaluate the depth and complexity of the character's personality development; the appearance dimension is assigned a 15% weight to evaluate the recognizability and completeness of the character's visual design; the social dimension is assigned a 15% weight to evaluate the richness and diversity of the character's relationship design; the functional dimension is assigned a 15% weight to evaluate the clarity and value of the game's functional positioning; and the narrative dimension is assigned a 15% weight to evaluate the depth and tension of the story's potential. Finally, a five-dimensional radar chart is obtained and output. The five-dimensional radar chart can intuitively display the morphological characteristics of the game characters and can be used for multi-character comparative analysis.

[0047] In specific embodiments, the character behavior evidence includes character behavior evidence under extreme situations, character behavior evidence under stress, character behavior evidence in daily life, and character behavior evidence based on surface information. Different weights are assigned to different types of character behavior evidence. Character behavior evidence under extreme situations is the behavior of the game character in life-or-death situations and core conflict scenarios, which best reveals the character's essence. Character behavior evidence under stress is the stress state of the game character in scenarios of being alone, fatigue, and emotional outburst. Character behavior evidence in daily life is the daily behavior of the game character under repetitive habits, stable patterns, and normal behaviors. Character behavior evidence based on surface information is the surface information of the character under polite behavior, professional behavior, and social mask.

[0048] In a specific embodiment, the large language model simulates the independent evaluation process of different cognitive subjects from three perspectives, calculates the cognitive evolution path, evaluates the game character settings from an NPC perspective based on externally visible information (such as appearance, identity, and initial behavior), adds gameplay experience information (plot dialogue, behavioral changes, and interactive feedback) to evaluate the player perspective based on the NPC evaluation, updates the score, and evaluates the designer perspective based on the aforementioned evaluation, combined with deep settings (background story, intrinsic motivation, and design intention), to reach the final score, and calculates the cognitive reversal degree of different perspectives, quantifying the degree of cognitive change from different perspectives. The calculation formula is as follows: ; Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings.

[0049] It should be noted that the domestic large language model is guided to perform professional evaluation through carefully designed prompt word templates: the large language model is positioned as a professional game character designer (15 years of experience) to ensure professionalism, the scoring criteria, weight configuration and evidence level system are encoded into a structured rule document, a fixed evaluation process is enforced to ensure the consistency of the evaluation process, the output format (table, radar chart and separator line) is strictly limited, and judgments beyond the scope are prohibited.

[0050] It should be noted that the professional workflow provided covers the entire process from character creation to evaluation, including: Complete evaluation workflow: Conduct a comprehensive multi-perspective evaluation of existing role settings; Assisted workflow for filling out forms: Guided creation of new character settings; Diagnose and optimize workflows: Identify logical problems and provide improvement suggestions; Format conversion workflow: Converts PDF / HTML source documents into standardized CSV format, including a three-stage verification mechanism and a five-level source labeling system; Source document review workflow: Perform a four-dimensional quality assessment on the source documents before conversion and generate a 5-star rating report.

[0051] In a specific implementation, a source document review workflow is performed on the PDF / HTML source documents to conduct a pre-assessment, checking the module integrity, content utilization, and structuring level; a format conversion workflow is performed to conduct format conversion and three-stage verification (basic data anchors, line-by-line verification of dialogue, and logical checks); and a decision is made on whether to continue the subsequent evaluation process based on the quality report.

[0052] It should be noted that this application converts subjective role settings into measurable values, ensuring the objectivity and comparability of evaluation results. Strict rules and standardized processes ensure that the consistency of evaluation results for the same role by different evaluators at different times exceeds 85%, and each score has a complete chain of evidence, supporting traceability and quality verification. Multi-perspective cognitive analysis is conducive to identifying the effectiveness of role reversal design, and the two-layer completeness assessment can accurately locate the type and priority of missing materials, reducing the blindness of material supplementation by more than 60%. The natural language understanding capability of the large language model can automatically complete complex tasks such as evidence extraction, logical reasoning, and score calculation, improving evaluation efficiency by 3-5 times. Through source document review and implementation log recording, a reusable evaluation knowledge base is formed, which can support the team's experience accumulation.

[0053] The following is for reference. Figure 3 It shows a schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application. Figure 3 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0054] like Figure 3 As shown, the computer system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 302 or programs loaded from storage section 308 into random access memory (RAM) 303. The RAM 303 also stores various programs and data required for system operation. The CPU 301, ROM 302, and RAM 303 are interconnected via bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0055] The following components are connected to I / O interface 305: an input section 306 including a keyboard, mouse, etc.; an output section 307 including a liquid crystal display (LCD) and speakers, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN card and a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to I / O interface 305 as needed. A removable medium 311, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on drive 310 as needed so that computer programs read from it can be installed into storage section 308 as needed.

[0056] Specifically, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 309, and / or installed from removable medium 311. When the computer program is executed by central processing unit (CPU) 301, it performs the functions defined in the methods of this application. It should be noted that the computer-readable storage medium of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having 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 thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. Program code contained on a computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0057] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0058] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0059] The modules described in the embodiments of this application can be implemented in software or in hardware.

[0060] In another aspect, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: inject the rules of the quantitative evaluation framework into a large language model through a preset prompt word template, and input a game character setting CSV file into the large language model; the large language model performs a comprehensive completeness evaluation of the game character setting based on a two-layer completeness evaluation system, and performs quantitative analysis of character personality traits through three-dimensional coordinates; the large language model performs a comprehensive quantitative evaluation of the game character setting based on a five-dimensional evaluation model, obtaining a five-dimensional radar chart, and assigning different weight coefficients to character behavior evidence of different credibility levels and performing quantitative analysis; the large language model performs NPC perspective evaluation, player perspective evaluation, and designer perspective evaluation of the game character setting, and calculates the cognitive reversal degree of different perspectives; the large language model performs automated reasoning and calculation, and outputs an evaluation report including comprehensive completeness, quantitative evaluation analysis results, cognitive reversal degree, and a five-dimensional radar chart.

[0061] Finally, it should be noted that the above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A quantitative evaluation method for game character design based on a large language model, characterized in that, include: S1: Inject the rules of the quantitative evaluation framework into the large language model through the preset prompt word template, and input the game character setting CSV file into the large language model; S2: The large language model is based on a two-layer completeness assessment system to comprehensively assess the completeness of game character settings, and uses three-dimensional coordinates to quantitatively analyze character personality traits; S3: The large language model is based on a five-dimensional evaluation model to conduct a comprehensive quantitative evaluation of the game character settings, obtain a five-dimensional radar chart, and assign different weight coefficients to character behavior evidence with different credibility and conduct quantitative analysis. S4: The large language model evaluates the game character settings from NPC perspective, player perspective, and designer perspective, and calculates the cognitive reversal degree from different perspectives. The calculation formula is as follows: ； Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings. S5: The large language model performs automated reasoning and calculation, and outputs an evaluation report including comprehensive completeness, quantitative evaluation analysis results, cognitive reversal degree and five-dimensional radar chart.

2. The method according to claim 1, characterized in that, The two-tier integrity assessment system includes a hard integrity assessment based on objective indicators and a perspective integrity assessment based on the perspectives of different cognitive subjects. The objective metrics include total text volume, number of dialogues, and number of scenes; The perspective integrity assessment includes evaluating the appearance, identity, initial behavior, and surface personality of NPCs; evaluating the plot performance, in-depth dialogue, behavioral changes, and interactive feedback from the player's perspective; and evaluating the background story, intrinsic motivation, design intentions, and deep settings from the designer's perspective.

3. The method according to claim 1, characterized in that, The formula for calculating the overall integrity assessment is as follows: S0=W1 S1+W2 S2+W3 S3+W4 S4; Wherein, S0 represents the overall completeness score, S1 represents the hard completeness score, S2 represents the NPC perspective score, S3 represents the player perspective score, S4 represents the designer perspective score, and W1, W2, W3 and W4 represent the weights of the hard completeness score, the NPC perspective score, the player perspective score, and the designer perspective score, respectively.

4. The method according to claim 1, characterized in that, The three-dimensional coordinate system includes a first coordinate axis representing the character's way of thinking, a second coordinate axis representing the character's emotional temperature, and a third coordinate axis representing the intensity of the character's performance. Each coordinate axis is divided into several numerical intervals. Combining the numerical intervals of each coordinate axis forms multiple personality positioning types.

5. The method according to claim 1, characterized in that, The five-dimensional evaluation model includes psychological, appearance, social, functional, and narrative dimensions, assigning different weights to different dimensions to evaluate the quality of game characters; The five-dimensional evaluation model assesses the depth and complexity of game character personality development through the psychological dimension, the recognizability and completeness of game character visual design through the appearance dimension, the richness and diversity of game character relationship design through the social dimension, the clarity and value of game function positioning through the functional dimension, and the depth and tension of story potential through the narrative dimension.

6. The method according to claim 1, characterized in that, The role behavior evidence includes role behavior evidence under extreme situations, role behavior evidence under stress, role behavior evidence in daily performance, and role behavior evidence based on surface information.

7. The method according to claim 6, characterized in that, Evidence of character behavior in extreme situations refers to the behavior of game characters in life-or-death situations and core conflict scenarios. Evidence of character behavior under stress refers to the stress state of game characters in situations of being alone in reality, fatigue reactions, and emotional outbursts. Evidence of character behavior in daily performance refers to the daily performance of game characters under repetitive habits, stable patterns, and normal behaviors. Evidence of character behavior in surface information refers to the surface information of characters under polite behavior, professional behavior, and social masks.

8. A quantitative evaluation system for game character design based on a large language model, characterized in that, include: Preset module: The rules of the quantitative evaluation framework are injected into the large language model through preset prompt word templates, and the game character setting CSV file is input into the large language model; First analysis module: The large language model uses a two-layer completeness assessment system to comprehensively evaluate the completeness of game character settings and uses three-dimensional coordinates to quantitatively analyze character personality traits; The second analysis module: The large language model performs a comprehensive quantitative evaluation of the game character settings based on the five-dimensional evaluation model, obtains a five-dimensional radar chart, and assigns different weight coefficients to character behavior evidence with different credibility and performs quantitative analysis. The third analysis module: The large language model evaluates the game character settings from NPC perspective, player perspective, and designer perspective, and calculates the cognitive reversal degree from different perspectives. The calculation formula is as follows: ； Wherein, IR represents cognitive reversal, X1 and X2 represent NPC perspective ratings, Y1 and Y2 represent player perspective ratings, and Z1 and Z2 represent designer perspective ratings. Output module: The large language model performs automated reasoning and calculation, and outputs an evaluation report including comprehensive completeness, quantitative evaluation analysis, cognitive reversal degree and five-dimensional radar chart.

9. The system according to claim 8, characterized in that, The system executes specific tasks according to a variety of preset workflow modes, including: a complete evaluation workflow for fully evaluating existing roles, an auxiliary filling workflow to guide the creation of new roles, a diagnostic optimization workflow to identify logical problems and provide improvement suggestions, a format conversion workflow to convert unstructured source documents into standardized formats, and a source document review workflow to assess the quality of the source documents before conversion and generate a rating report.

10. A computer program product having one or more computer programs thereon, characterized in that, When the computer program is executed by a computer processor, the method described in any one of claims 1-7 is performed.

Images