Patience-Discounted Reward Evaluation on NoveltyBench
4.096UtilityQwen3-14B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-14BCalibration=Soft2026.05 | 4.096 | |
| Qwen3-14BCalibration=Hard2026.05 | 3.849 | |
| Qwen3-8BCalibration=Soft2026.05 | 3.843 | |
| GPT-OSS-20BCalibration=Base2026.05 | 3.786 | |
| Qwen3-8BCalibration=Hard2026.05 | 3.734 | |
| Qwen3-4BCalibration=Soft2026.05 | 3.559 | |
| Qwen3-14BCalibration=Base2026.05 | 3.392 | |
| Llama-3.2-3B-itCalibration=Base2026.05 | 3.385 | |
| Qwen3-8BCalibration=Base2026.05 | 3.378 | |
| Qwen3-4BCalibration=Hard2026.05 | 3.322 | |
| Llama-3.2-3B-itCalibration=Hard2026.05 | 3.183 | |
| Qwen3-1.7BCalibration=Soft2026.05 | 3.143 | |
| Llama-3.2-3B-itCalibration=Soft2026.05 | 3.126 | |
| Qwen3-1.7BCalibration=Hard2026.05 | 3.094 | |
| Gemma-3-4B-itCalibration=Soft2026.05 | 3.034 | |
| Qwen3-4BCalibration=Base2026.05 | 2.915 | |
| Gemma-3-12B-itCalibration=Soft2026.05 | 2.863 | |
| Gemma-3-27B-itCalibration=Soft2026.05 | 2.858 | |
| Qwen3-1.7BCalibration=Base2026.05 | 2.804 | |
| Gemma-3-4B-itCalibration=Hard2026.05 | 2.776 | |
| Gemma-3-12B-itCalibration=Hard2026.05 | 2.678 | |
| Qwen3-0.6BCalibration=Base2026.05 | 2.639 | |
| Gemma-3-27B-itCalibration=Hard2026.05 | 2.623 | |
| GPT-OSS-20BCalibration=Hard2026.05 | 2.615 | |
| Llama-3.2-1B-itCalibration=Base2026.05 | 2.56 | |
| Gemma-3-1B-itCalibration=Soft2026.05 | 2.555 | |
| Gemma-3-1B-itCalibration=Hard2026.05 | 2.527 | |
| Llama-3.2-1B-itCalibration=Hard2026.05 | 2.466 | |
| Llama-3.2-1B-itCalibration=Soft2026.05 | 2.388 | |
| Gemma-3-4B-itCalibration=Base2026.05 | 2.332 | |
| Gemma-3-1B-itCalibration=Base2026.05 | 2.303 | |
| Gemma-3-27B-itCalibration=Base2026.05 | 2.27 | |
| GPT-OSS-20BCalibration=Soft2026.05 | 2.179 | |
| Gemma-3-12B-itCalibration=Base2026.05 | 2.173 | |
| Qwen3-0.6BCalibration=Hard2026.05 | 2.04 | |
| Qwen3-0.6BCalibration=Soft2026.05 | 1.654 |