Commonsense Question Answering on ARC-C
41.13AccuracyPHSA
Evaluation results
| Method | Links | |
|---|---|---|
| PHSATraining Tokens=20B, Inference Top-K=22026.01 | 41.13 | |
| PHSATraining Tokens=20B, Inference Top-K=N/A2026.01 | 40.87 | |
| DenseTraining Tokens=20B, Inference Top-K=N/A2026.01 | 40.44 | |
| DenseTraining Tokens=20B, Inference Top-K=22026.01 | 40.36 | |
| InfLLM v2Training Tokens=20B, Inference Top-K=22026.01 | 39.51 | |
| InfLLM v2Training Tokens=20B, Inference Top-K=N/A2026.01 | 39.33 | |
| Qwen-0.6B-BaseTraining Tokens=N/A, Inference Top-K=N/A2026.01 | 33.45 | |
| Qwen-0.6B-BaseTraining Tokens=N/A, Inference Top-K=22026.01 | 33.28 |