Factuality Evaluation on BIO (test)
88.9FS ScoreCaLF
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CaLFBase Model=MistralOrca-7B, Retrieval Filtering=true2024.06 | 88.9 | 92.7 | |
| Few-shot FTBase Model=MistralOrca-7B, Retrieval Filtering=true2024.06 | 86.1 | 69.3 | |
| CaLFBase Model=MistralOrca-7B2024.06 | 83.4 | 92.7 | |
| Self-RAGParameter Count=7B2024.06 | 81.2 | — | |
| Self-RAGParameter Count=13B2024.06 | 80.2 | — | |
| Llama 2Parameter Count=13B, Mode=chat2024.06 | 79.9 | — | |
| Few-shot FTBase Model=MistralOrca-7B2024.06 | 78.7 | 69.3 | |
| ChatGPTRetrieval=w/o retrieval2024.06 | 71.8 | — |