Evidence Retrieval on FAKEWIKI NoiseInjection
63.5Recall@10SCORINGMODEL
Evaluation Results
| Method | Links | |
|---|---|---|
| SCORINGMODELTarget Model=Llama-3.1-8B2026.05 | 63.5 | |
| SCORINGMODELTarget Model=Qwen3-8B2026.05 | 62.1 | |
| STEERFUSETarget Model=Qwen3-8B2026.05 | 51.1 | |
| STEERFUSETarget Model=Llama-3.1-8B2026.05 | 38.1 | |
| ContrieverTarget Model=Qwen3-8B, Query Type=answer2026.05 | 26.6 | |
| BGE-baseTarget Model=Qwen3-8B, Query Type=answer2026.05 | 25.7 | |
| SBERT-MiniLMTarget Model=Qwen3-8B, Query Type=QA2026.05 | 25 | |
| SBERT-MiniLMTarget Model=Qwen3-8B, Query Type=answer2026.05 | 23.9 | |
| BGE-baseTarget Model=Qwen3-8B, Query Type=QA2026.05 | 23.6 | |
| Finetuned EmbedSimTarget Model=Llama-3.1-8B2026.05 | 23.1 | |
| SBERT-MPNetTarget Model=Qwen3-8B, Query Type=answer2026.05 | 20.1 | |
| SBERT-MiniLMTarget Model=Llama-3.1-8B, Query Type=QA2026.05 | 18.5 | |
| Finetuned EmbedSimTarget Model=Qwen3-8B2026.05 | 16.4 | |
| ContrieverTarget Model=Qwen3-8B, Query Type=QA2026.05 | 15 | |
| SBERT-MPNetTarget Model=Qwen3-8B, Query Type=QA2026.05 | 14.8 | |
| BGE-baseTarget Model=Llama-3.1-8B, Query Type=QA2026.05 | 13.8 | |
| BGE-baseTarget Model=Llama-3.1-8B, Query Type=answer2026.05 | 13.1 | |
| SBERT-MiniLMTarget Model=Llama-3.1-8B, Query Type=answer2026.05 | 12.9 | |
| ContrieverTarget Model=Llama-3.1-8B, Query Type=answer2026.05 | 12.5 | |
| SBERT-MPNetTarget Model=Llama-3.1-8B, Query Type=answer2026.05 | 11.6 | |
| SBERT-MPNetTarget Model=Llama-3.1-8B, Query Type=QA2026.05 | 9.8 | |
| ContrieverTarget Model=Llama-3.1-8B, Query Type=QA2026.05 | 7.1 | |
| MinHashTarget Model=Qwen3-8B, Query Type=answer2026.05 | 0.1 | |
| MinHashTarget Model=Qwen3-8B, Query Type=QA2026.05 | 0.1 | |
| MinHashTarget Model=Llama-3.1-8B, Query Type=answer2026.05 | 0.1 | |
| MinHashTarget Model=Llama-3.1-8B, Query Type=QA2026.05 | 0.1 |