| { | |
| "model": "Qev-2B", | |
| "method": "Teacher-probability cross entropy followed by replay and representation-response distillation", | |
| "candidate_preview": false, | |
| "seed": 17, | |
| "continuation_steps": 800, | |
| "full_suites": { | |
| "decision_dev": { | |
| "qev_2b": { | |
| "questions": 1468, | |
| "correct": 1266, | |
| "accuracy": 0.8623978201634878 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 1468, | |
| "correct": 927, | |
| "accuracy": 0.6314713896457765 | |
| } | |
| }, | |
| "transfer_dev": { | |
| "qev_2b": { | |
| "questions": 764, | |
| "correct": 584, | |
| "accuracy": 0.7643979057591623 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 764, | |
| "correct": 493, | |
| "accuracy": 0.6452879581151832 | |
| } | |
| }, | |
| "mmlupro": { | |
| "qev_2b": { | |
| "questions": 1000, | |
| "correct": 387, | |
| "accuracy": 0.387 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 1000, | |
| "correct": 312, | |
| "accuracy": 0.312 | |
| } | |
| }, | |
| "semif": { | |
| "qev_2b": { | |
| "questions": 252, | |
| "correct": 215, | |
| "accuracy": 0.8531746031746031 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 252, | |
| "correct": 152, | |
| "accuracy": 0.6031746031746031 | |
| } | |
| }, | |
| "scienthoon": { | |
| "qev_2b": { | |
| "questions": 873, | |
| "correct": 628, | |
| "accuracy": 0.7193585337915235 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 873, | |
| "correct": 470, | |
| "accuracy": 0.5383734249713631 | |
| } | |
| }, | |
| "wanli": { | |
| "qev_2b": { | |
| "questions": 256, | |
| "correct": 173, | |
| "accuracy": 0.67578125 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 256, | |
| "correct": 129, | |
| "accuracy": 0.50390625 | |
| } | |
| }, | |
| "public": { | |
| "qev_2b": { | |
| "questions": 231, | |
| "correct": 172, | |
| "accuracy": 0.7445887445887446 | |
| }, | |
| "qwen35_2b_base": { | |
| "questions": 231, | |
| "correct": 146, | |
| "accuracy": 0.6320346320346321 | |
| } | |
| } | |
| } | |
| } | |