Psychological Test LLM Training With RAG Context Templates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing large language models (LLMs) are not effectively utilized in psychology and motivation science due to limitations in data volume, technology, and inability to understand individual/group psychological conditions, leading to inaccurate user testing results.
Innovation Solution
A cloud-based big data and AI platform, Motivation Quotient (MQ), integrates personal psychological test data using retrieval augmented generation (RAG) to analyze motivational factors, providing tailored responses through a trained LLM engine with pre-defined context templates and system commands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If generic AI engines like ChatGPT are used for user testing, then the system can provide quick responses, but the accuracy and reliability of psychological analysis deteriorates due to hallucinations
Solution Approach 1:
The patent introduces a specialized LLM trained on psychological test data as an intermediary between the generic AI engine and the psychological analysis task. This specialized model acts as a mediator that filters out hallucinations while maintaining response efficiency, resolving the contradiction between speed and accuracy.
Solution Approach 2:
The system changes the parameter of model specialization by transitioning from a generic AI engine to a specifically trained LLM on psychological test data. This parameter change enables the system to maintain both quick responses and high accuracy by optimizing the model for psychological analysis tasks.
2Device complexity
If LLMs are used without specialized training data, then the system remains simple and fast to deploy, but the ability to understand individual psychological conditions deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-training the LLM on psychological test data before deployment. This preliminary training equips the model with the necessary psychological knowledge and analysis capabilities, enabling it to understand individual psychological conditions without adding complexity to the system architecture.
Solution Approach 2:
The specialized LLM serves multiple functions: it processes psychological test data, analyzes motivational factors, and provides personalized insights. This multi-functionality achieves high adaptability while maintaining system simplicity by consolidating multiple tasks into a single trained model.
3Speed
If existing LLMs are used for motivational analysis, then the system can process data quickly, but the measurement precision of motivational factors deteriorates due to lack of specialized knowledge
Solution Approach 1:
The system changes the parameter of model specialization by training the LLM on motivational psychology data. This enables the model to maintain fast processing speeds while significantly improving measurement precision in analyzing motivational factors through specialized knowledge of psychological theory and terminology.
Data Source
AI summary
A method for a web platform to train a large language model platform (LLM) to respond to user inquiries, which includes obtaining a context-data from a data analysis engine, determining a context based on the context-data, selecting a pre-trained context template for the context, determining a system command for the LLM platform, and transmitting the context-data, the pre-trained context template, and the system command to the LLM platform.