General Knowledge Reasoning on MMLU-Pro (test)
37.72AccuracyIntroLLM
Evaluation Results
| Method | Links | |
|---|---|---|
| IntroLLMBackbone=Qwen3-4B-Base2026.02 | 37.72 | |
| TampoBackbone=Qwen3-4B-Base, temperature range=0.6-1.52026.02 | 37.23 | |
| GRPOBackbone=Qwen3-4B-Base, temperature (τ)=1.02026.02 | 36.6 | |
| EADBackbone=Qwen3-4B-Base, annealing range=1.2→0.12026.02 | 36.22 | |
| IntroLLMBackbone=Qwen3-1.7B-Base2026.02 | 28.89 | |
| EADBackbone=Qwen3-1.7B-Base, annealing range=1.2→0.12026.02 | 28.66 | |
| GRPOBackbone=Qwen3-1.7B-Base, temperature (τ)=1.02026.02 | 28.39 | |
| TampoBackbone=Qwen3-1.7B-Base, temperature range=0.6-1.52026.02 | 27.16 | |
| BaseBackbone=Qwen3-4B-Base2026.02 | 25.06 | |
| BaseBackbone=Qwen3-1.7B-Base2026.02 | 15.49 |