Comprehensive cognitive reasoning on MMLU
53.43MMLU ScorePHSA
Evaluation Results
| Method | Links | |
|---|---|---|
| PHSATraining Tokens=20B, Inference Top-K=N/A2026.01 | 53.43 | |
| DenseTraining Tokens=20B, Inference Top-K=N/A2026.01 | 53.42 | |
| PHSATraining Tokens=20B, Inference Top-K=22026.01 | 53.17 | |
| DenseTraining Tokens=20B, Inference Top-K=22026.01 | 53.09 | |
| InfLLM v2Training Tokens=20B, Inference Top-K=N/A2026.01 | 52.95 | |
| Qwen-0.6B-BaseTraining Tokens=N/A, Inference Top-K=N/A2026.01 | 52.69 | |
| InfLLM v2Training Tokens=20B, Inference Top-K=22026.01 | 52.68 | |
| Qwen-0.6B-BaseTraining Tokens=N/A, Inference Top-K=22026.01 | 51.91 |