Psychological assessment method based on large language model Agent
By introducing multiple large language models and knowledge graphs that load specific prompt words into the psychological assessment method, the problems of context information loss and insufficient robustness of the confrontation sample are solved, efficient and personalized psychological assessment is achieved, and resource requirements are reduced.
Patent Information
- Application Number
- CN202411900929.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-23
- Publication Date
- 2025-04-15
AI Technical Summary
The existing psychological evaluation method based on large language models is lost when processing long texts and multiple rounds of conversations, and the adversarial samples are insufficiently robust, resource requirements are high, making it difficult to effectively apply in resource-limited environments.
A psychological evaluation method based on the large language model Agent is proposed, including the user psychological feature extraction module, the user psychological state search module, the agent questioning module and the psychological evaluation report generation module. Through multiple large language models loading specific prompt words and connecting them through thinking chains, maintaining context consistency; introducing knowledge graphs and vector knowledge bases to improve the robustness of the adversarial samples, and optimizing resource utilization through modular design.
It enhances context retention capabilities, improves the robustness of the countermeasures, reduces resource requirements, realizes personalized and precise psychological assessments, and improves the effectiveness of user experience and evaluation results.