Belief assessment on misinformation on MISBELIEF Health domain Hard
3.63Belief ScoreQwen2.5-72B
Evaluation results
| Method | Links | |
|---|---|---|
| Qwen2.5-72BSetting=Round 32026.01 | 3.63 | |
| GPT-3.5-turboSetting=Round 32026.01 | 3.5 | |
| Qwen2.5-72BSetting=Round 12026.01 | 3.46 | |
| Qwen-turboSetting=Round 32026.01 | 3.44 | |
| Qwen2.5-32BSetting=Round 32026.01 | 3.26 | |
| GPT-5Setting=Round 32026.01 | 3.21 | |
| GPT-3.5-turboSetting=Round 12026.01 | 3.19 | |
| Qwen-turboSetting=Round 12026.01 | 3.12 | |
| Qwen2.5-32BSetting=Round 12026.01 | 3.08 | |
| GPT-5Setting=Round 12026.01 | 2.95 | |
| Llama3-8BSetting=Round 32026.01 | 2.94 | |
| Qwen-turbo (w/o reasoning)Setting=Round 32026.01 | 2.88 | |
| Llama3-8BSetting=Round 12026.01 | 2.71 | |
| Qwen-turbo (w/o reasoning)Setting=Round 12026.01 | 2.69 | |
| Qwen2.5-72BSetting=Baseline2026.01 | 2.38 | |
| Qwen2.5-72BSetting=Original2026.01 | 2.35 | |
| Qwen-turbo (w/o reasoning)Setting=Baseline2026.01 | 2 | |
| GPT-3.5-turboSetting=Baseline2026.01 | 1.99 | |
| GPT-3.5-turboSetting=Original2026.01 | 1.97 | |
| Qwen-turbo (w/o reasoning)Setting=Original2026.01 | 1.95 | |
| Qwen-turboSetting=Original2026.01 | 1.9 | |
| Llama3-8BSetting=Original2026.01 | 1.82 | |
| Qwen-turboSetting=Baseline2026.01 | 1.8 | |
| Qwen2.5-32BSetting=Original2026.01 | 1.77 | |
| GPT-5Setting=Original2026.01 | 1.67 | |
| Qwen2.5-32BSetting=Baseline2026.01 | 1.6 | |
| GPT-5Setting=Baseline2026.01 | 1.51 | |
| Llama3-8BSetting=Baseline2026.01 | 1.51 |