Factuality Correction on BIO (test)
51PrecisionRAC
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| RACBase Model=gpt-oss-120b2026.01 | 51 | 28 | 42 | 11 | 28 | |
| RACBase Model=gpt-oss-120b2026.01 | 51 | 28 | 42 | 11 | 28 | |
| LLM2Base Model=gpt-oss-120b2026.01 | 48 | 31 | 41 | 10 | 27 | |
| LLM2Base Model=gpt-oss-120b2026.01 | 48 | 31 | 41 | 10 | 27 | |
| FCBase Model=mixtral-8x22b-instruct2026.01 | 46 | 33 | 41 | 12 | 11 | |
| FACTCORRECTORBase Model=mixtral-8x22b-instruct2026.01 | 46 | 33 | 41 | 12 | 11 | |
| LLM2Base Model=llama-3.3-70b-instruct2026.01 | 45 | 21 | 36 | 12 | 18 | |
| LLM2Base Model=llama-3.3-70b-instruct2026.01 | 45 | 21 | 36 | 12 | 18 | |
| FCBase Model=llama-3.3-70b-instruct2026.01 | 44 | 22 | 35 | 12 | 17 | |
| FACTCORRECTORBase Model=llama-3.3-70b-instruct2026.01 | 44 | 22 | 35 | 12 | 17 | |
| LLM2Base Model=mixtral-8x22b-instruct2026.01 | 42 | 30 | 37 | 11 | 7 | |
| RACBase Model=llama-3.3-70b-instruct2026.01 | 42 | 21 | 33 | 12 | 17 | |
| LLM2Base Model=mixtral-8x22b-instruct2026.01 | 42 | 30 | 37 | 11 | 7 | |
| RACBase Model=llama-3.3-70b-instruct2026.01 | 42 | 21 | 33 | 12 | 16 | |
| FCBase Model=gpt-oss-120b2026.01 | 40 | 29 | 36 | 10 | 24 | |
| FACTCORRECTORBase Model=gpt-oss-120b2026.01 | 40 | 29 | 36 | 10 | 24 | |
| RACBase Model=mixtral-8x22b-instruct2026.01 | 39 | 29 | 34 | 5 | 13 | |
| RACBase Model=mixtral-8x22b-instruct2026.01 | 39 | 29 | 34 | 5 | 13 | |
| CRITICBase Model=llama-3.3-70b-instruct2026.01 | 37 | 18 | 29 | 8 | 13 | |
| CRITICBase Model=llama-3.3-70b-instruct2026.01 | 37 | 18 | 29 | 8 | 13 | |
| CRITICBase Model=mixtral-8x22b-instruct2026.01 | 27 | 22 | 25 | 5 | 5 | |
| LLM1Base Model=mixtral-8x22b-instruct2026.01 | 27 | 25 | 27 | 2 | 7 | |
| CRITICBase Model=mixtral-8x22b-instruct2026.01 | 27 | 22 | 25 | 5 | 5 | |
| LLM1Base Model=mixtral-8x22b-instruct2026.01 | 27 | 25 | 27 | 2 | 7 | |
| FCBase Model=granite-4.0-h-small2026.01 | 26 | 10 | 20 | 14 | 3 | |
| FACTCORRECTORBase Model=granite-4.0-h-small2026.01 | 26 | 10 | 20 | 14 | 3 | |
| LLM1Base Model=llama-3.3-70b-instruct2026.01 | 23 | 14 | 19 | 7 | 11 | |
| LLM1Base Model=llama-3.3-70b-instruct2026.01 | 23 | 14 | 19 | 7 | 11 | |
| RACBase Model=granite-4.0-h-small2026.01 | 22 | 8 | 16 | 10 | 4 | |
| RACBase Model=granite-4.0-h-small2026.01 | 22 | 8 | 16 | 10 | 4 | |
| LLM2Base Model=granite-4.0-h-small2026.01 | 19 | 13 | 16 | 13 | 3 | |
| LLM2Base Model=granite-4.0-h-small2026.01 | 19 | 13 | 16 | 13 | 3 | |
| CRITICBase Model=granite-4.0-h-small2026.01 | 12 | 2 | 8 | 4 | 2 | |
| CRITICBase Model=granite-4.0-h-small2026.01 | 12 | 2 | 8 | 4 | 2 | |
| SFTBase Model=mixtral-8x22b-instruct2026.01 | 5 | 0 | 2 | 5 | 2 | |
| SFTBase Model=granite-4.0-h-small2026.01 | 5 | 0 | 2 | 6 | -2 | |
| SFTBase Model=llama-3.3-70b-instruct2026.01 | 1 | -2 | -1 | 6 | -3 | |
| CRITICBase Model=gpt-oss-120b2026.01 | 0 | 0 | 0 | 0 | 1 | |
| CRITICBase Model=gpt-oss-120b2026.01 | 0 | 0 | 0 | 0 | 1 | |
| LLM1Base Model=granite-4.0-h-small2026.01 | -3 | -1 | -2 | 2 | 0 | |
| LLM1Base Model=granite-4.0-h-small2026.01 | -3 | -1 | -2 | 2 | 0 | |
| SFTBase Model=gpt-oss-120b2026.01 | -4 | 0 | -2 | 0 | -4 | |
| LLM1Base Model=gpt-oss-120b2026.01 | -28 | -10 | -22 | -24 | 6 | |
| LLM1Base Model=gpt-oss-120b2026.01 | -28 | -10 | -22 | -24 | 6 |