Certainty Distortion Evaluation on Academic Papers document-level
15CDgpt-5-nano
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| gpt-5-nano2026.06 | 15 | 11.3 | 3.1 | |
| Llama 3.3parameters=70B2026.06 | 15.8 | 12.3 | 12.3 | |
| Qwen3parameters=8B2026.06 | 18.4 | 15.8 | 6.3 | |
| Claude 4.5variant=H2026.06 | 22.8 | 21 | 11.3 | |
| Gemini 3.1variant=fl2026.06 | 24.8 | 22.7 | 10.8 |