728x90
반응형
The advisor embodies two key LLM-as-a-Judge concepts:
- Point-wise Evaluation: Each response receives an individual quality score based on predefined criteriaSelf
- Refinement: Failed responses trigger retry attempts with constructive feedback to guide improvement
https://huggingface.co/learn/cookbook/en/llm_judge#3-improve-the-llm-judge
728x90
'07.AI' 카테고리의 다른 글
| LLM - 성능 - 벤치마크 - Extended NYT Connections (0) | 2025.12.12 |
|---|---|
| LLM - 검색 증강 생성 (RAG, Retrieval-augmented generation) (3) | 2025.12.06 |
| 생성형 AI - 소버린 AI(Sovereign AI) (5) | 2025.12.02 |
| 머신러닝 - 강화학습 - GRPO 알고리즘 (0) | 2025.12.01 |
| 인공지능 - 위험관리 - AssurAI 데이터셋 (0) | 2025.11.25 |
