25 May 2023
주안점 self-provided feedback
generate an initial output using an LLM; then, the same LLM provides feedback for its output and uses it to refine itself, iteratively
until a stopping condition is met. The stopping condition stop(f bt, t) either stops at a specified timestep t, or extracts a stopping indicator (e.g. a scalar stop score) from the feedback.
To inform the model about the previous iterations, we retain the history of previous feedback and outputs by appending them to the prompt
'프롬프팅' 카테고리의 다른 글
| SELF-CONTRADICTORY HALLUCINATIONS OF LLMS:EVALUATION, DETECTION AND MITIGATION 논문리뷰 (0) | 2024.08.08 |
|---|---|
| The Prompt Report: A Systematic Survey of Prompting Techniques 논문 리뷰 (프롬프팅 기법에 관해) (0) | 2024.07.22 |