Fact Verification on FactKG
67.44AccuracySeleCom
Evaluation results
| Method | Links | |
|---|---|---|
| SeleComBackbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=82x2026.01 | 67.44 | |
| SeleComBackbone=Mistral-7B-Instruct, Top-k=5, Comp. Rate=82x2026.01 | 66.5 | |
| PISCOBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=64x2026.01 | 65.78 | |
| SeleComBackbone=Qwen2.5-7B-Instruct, Top-k=5, Comp. Rate=82x2026.01 | 65.24 | |
| SeleComBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=82x2026.01 | 65.23 | |
| LLM (with RAG*)Backbone=Mistral-7B-Instruct, Top-k=5, Comp. Rate=-2026.01 | 65.1 | |
| LLM (with RAG*)Backbone=Qwen2.5-7B-Instruct, Top-k=5, Comp. Rate=-2026.01 | 64.6 | |
| LLM (with RAG)Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=-2026.01 | 64.16 | |
| xRAGBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=164x2026.01 | 63 | |
| ICAEBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=4x2026.01 | 62.7 | |
| LLM (with RAG*)Backbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=-2026.01 | 62.6 | |
| LLM (with RAG)Backbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=-2026.01 | 62.3 | |
| LLMLingua-2Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=4x2026.01 | 61.65 | |
| LLM (with RAG*)Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=-2026.01 | 61.23 | |
| COCOMBackbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=128x2026.01 | 60.87 | |
| LLM (w/o RAG)Backbone=Mistral-7B-Instruct, Top-k=1, Comp. Rate=-2026.01 | 58.34 | |
| LLM (w/o RAG)Backbone=Qwen2.5-7B-Instruct, Top-k=1, Comp. Rate=-2026.01 | 57.93 |