
ICSFSurvey
In depth study of internal consistency and self feedback in large language models
- Provide a theoretical framework for internal consistency to explain reasoning gaps and hallucinations in LLMs.
- Self feedback mechanisms, including self-evaluation and self-renewal, to enhance model response or the model itself.
- Systematic classification research, categorizing tasks and work lines based on self feedback mechanisms.
- Summarize evaluation methods and benchmark tests to assess the effectiveness of self feedback.
- Explore key points such as whether self feedback is truly effective, and propose hypotheses such as the hourglass evolution of internal consistency.
- Overview of future research directions on internal consistency and self feedback in LLMs.
Product Details
ICSFSurvey is a survey study on the internal consistency and self feedback of large language models. It provides a unified perspective on the self-assessment and self-renewal mechanisms of LLMs, including theoretical frameworks, systematic classifications, evaluation methods, future research directions, and more.