Claim Correction on FEVER Retrieved evidence
50.7141SARI (%)M2C+
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| M2C+LLM=Qwen-2.5 (32B), Ranker=n/a, Masking=Diversity-aware (DM)2026.04 | 50.7141 | -2.5724 | |
| M2CLLM=Qwen-2.5 (32B), Ranker=MonoT5, Masking=Diversity-aware (DM)2026.04 | 50.3638 | -2.6848 | |
| M2Cw/VerLLM=Qwen-2.5 (32B), Ranker=MonoT52026.04 | 49.8118 | -3.0608 | |
| M2C+LLM=Llama-2.0 (70B), Ranker=n/a, Masking=Diversity-aware (DM)2026.04 | 49.3749 | -2.5939 | |
| M2CLLM=Qwen-2.5 (32B), Ranker=Contriever, Masking=Diversity-aware (DM)2026.04 | 49.285 | -2.6159 | |
| M2CLLM=Llama-2.0 (70B), Ranker=MonoT5, Masking=Diversity-aware (DM)2026.04 | 49.1997 | -2.6702 | |
| M2CLLM=Llama-2.0 (70B), Ranker=Contriever, Masking=Diversity-aware (DM)2026.04 | 48.6495 | -2.6125 | |
| M2CLLM=Qwen-2.5 (32B), Ranker=ColBERT, Masking=Diversity-aware (DM)2026.04 | 48.3629 | -2.6326 | |
| M2Cw/VerLLM=Llama-2.0 (70B), Ranker=MonoT52026.04 | 48.2327 | -2.6143 | |
| M2CLLM=Llama-2.0 (70B), Ranker=ColBERT, Masking=Diversity-aware (DM)2026.04 | 47.4645 | -2.6491 | |
| RAGLLM=Qwen-2.5 (32B), Ranker=MonoT52026.04 | 46.9218 | -2.759 | |
| M2CLLM=Qwen-2.5 (32B), Ranker=RM3, Masking=Diversity-aware (DM)2026.04 | 46.5141 | -2.6143 | |
| M2CLLM=Qwen-2.5 (32B), Ranker=BM25, Masking=Diversity-aware (DM)2026.04 | 45.9969 | -2.7278 | |
| M2CLLM=Llama-2.0 (70B), Ranker=RM3, Masking=Diversity-aware (DM)2026.04 | 45.5321 | -2.6088 | |
| M2CLLM=Llama-2.0 (70B), Ranker=BM25, Masking=Diversity-aware (DM)2026.04 | 45.0705 | -2.5935 | |
| 0-shotLLM=Llama-2.0 (70B), Ranker=n/a2026.04 | 44.9646 | -3.7971 | |
| 0-shotLLM=Qwen-2.5 (32B), Ranker=n/a2026.04 | 43.2337 | -2.7315 | |
| RAGLLM=Llama-2.0 (70B), Ranker=MonoT52026.04 | 41.7222 | -2.3939 | |
| ZeroFeC-DALLM=n/a, Predictor=ZeroFeC-DA, Ranker=MonoT52026.04 | 39.3734 | -2.8629 | |
| T5-DistantLLM=n/a, Predictor=T5-Distant, Ranker=GENRE2026.04 | 34.9734 | -5.412 | |
| COMPEDITLLM=n/a, Predictor=COMPEDIT, Ranker=MonoT52026.04 | 26.0694 | -3.1801 |