rl4phyx-backup / eval_footprint /comparison_report.json
YUNTA88's picture
Upload folder using huggingface_hub
c6cbaf4 verified
Raw History Blame Contribute Delete
2.19 kB
{
"timestamp": "2026-02-26 04:38:24",
"scoring_method": "rule-based + Gemini 2.5 Flash judge (parallel)",
"base": {
"model": "Base",
"total": 1533,
"rule_correct": 48,
"gemini_rescued": 108,
"final_correct": 156,
"final_acc": 10.18,
"category_stats": {
"Mechanics": {
"total": 276,
"rule_correct": 8,
"gemini_correct": 33,
"final_correct": 41
},
"Waves/Acoustics": {
"total": 253,
"rule_correct": 7,
"gemini_correct": 20,
"final_correct": 27
},
"Electromagnetism": {
"total": 275,
"rule_correct": 7,
"gemini_correct": 16,
"final_correct": 23
},
"Modern Physics": {
"total": 222,
"rule_correct": 6,
"gemini_correct": 14,
"final_correct": 20
},
"Optics": {
"total": 252,
"rule_correct": 16,
"gemini_correct": 15,
"final_correct": 31
},
"Thermodynamics": {
"total": 255,
"rule_correct": 4,
"gemini_correct": 10,
"final_correct": 14
}
}
},
"sft": {
"model": "SFT",
"total": 1533,
"rule_correct": 41,
"gemini_rescued": 69,
"final_correct": 110,
"final_acc": 7.18,
"category_stats": {
"Mechanics": {
"total": 276,
"rule_correct": 2,
"gemini_correct": 27,
"final_correct": 29
},
"Waves/Acoustics": {
"total": 253,
"rule_correct": 8,
"gemini_correct": 12,
"final_correct": 20
},
"Electromagnetism": {
"total": 275,
"rule_correct": 7,
"gemini_correct": 8,
"final_correct": 15
},
"Modern Physics": {
"total": 222,
"rule_correct": 7,
"gemini_correct": 10,
"final_correct": 17
},
"Optics": {
"total": 252,
"rule_correct": 13,
"gemini_correct": 6,
"final_correct": 19
},
"Thermodynamics": {
"total": 255,
"rule_correct": 4,
"gemini_correct": 6,
"final_correct": 10
}
}
},
"improvement": "-3.00%"
}