논문 리뷰/Evaluation
-
ALCE (Automatic LLMs' Citation Evaluation)논문 리뷰/Evaluation 2024. 5. 22. 10:39
ALCE is a benchmark for Automatic LLMs' Citation Evaluation. ALCE collects a diverse set of questions and retrieval corpora and requires building end-to-end systems to retrieve supporting evidence and generate answers with citations. 평가 항목fluency, correctness, and citation quality- Use MAUVE (Pillutla et al., 2021) to measure fluency, adopt a natural language inference (NLI) model (Honovich et ..