Contextual Understanding and Reasoning on OpenHuEval
63.03HuWildBench WBScoreQwen3-4B
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Qwen3-4BImplementation patching=Patched2026.01 | 63.03 | 7.3 | 62.47 | 74.98 | 39.59 | 5.94 | 13.2 | 1.08 | 33.44 | |
| Racka-4BImplementation patching=Patched2026.01 | 57.17 | 10.05 | 61.94 | 77.53 | 38.93 | 4.68 | 18.98 | 2.15 | 33.93 | |
| Qwen3-4B-BaseImplementation patching=None2026.01 | 52.59 | 5.91 | 41.15 | 0 | 42.3 | 5.58 | 0 | 0 | 18.44 | |
| PULI-LlumiX-Llama-3.1 8BImplementation patching=None2026.01 | 17.77 | 20.03 | 75.86 | 77.36 | 33.54 | 3.96 | 29.16 | 2.15 | 32.47 |