Faithfulness detection on FaithCoT-Bench
61.7F1 ScoreGeoFaith
Evaluation results
| Method | Links | |
|---|---|---|
| GeoFaith2026.05 | 61.7 | |
| DeepSeek-V32026.05 | 59.1 | |
| GPT-o12026.05 | 56.3 | |
| Qwen2.5-32B-Instruct2026.05 | 52.3 | |
| LogicReward2026.05 | 50.4 | |
| GPT-4o2026.05 | 45.4 | |
| o3-mini2026.05 | 43.1 | |
| FaithLens2026.05 | 37.6 | |
| HHEM2.12026.05 | 37.5 | |
| Llama-3.1-70B-Instruct2026.05 | 32.7 |