https://openreview.net/pdf?id=RAtrnAtAsM
2. methodology
(1) Fine-grained World Knowledge Integration Module
(2) Iterative Feedback and Refinement Module
one LLM serve as Explainer - initial 결과물 생성
Critic LLM - Obeservation, Iterative feed back 받음 -> refine its explanation
2.1 Fine-grained World Knowledge Integration
inception prompt -> 에 의해
Cause Analyst role
Effect Analyst role
가 two LLMs에게 assign됨
The third agent, serves as the Knowledge Master for answering the questions of the reasoners, does not require a specific prompt.
After the roles are assigned, the Cause Analyst M and Effect Analyst A will collaborate in reasoning about the finegrained world knowledge by thought and action
Formally, we denote the Cause Analyst message obtained at time t by Mt and the corresponding observation from Knowledge Master is Omt , the Effect Analyst message is At and the corresponding observation is Oat , the historical messages
