med-record-audit / experiments /comparison.json
gauri-emergent
submission: final results, README rewrite, eval pipeline
b2d8977
Raw
History Blame Contribute Delete
792 Bytes
{
"agents_loaded": [
"Random",
"Naive LLM",
"Smart LLM",
"Trained"
],
"agents_missing": [],
"rows": [
{
"case": "easy_001",
"difficulty": "easy",
"Random": 0.2222,
"Naive LLM": 0.01,
"Smart LLM": 0.7923,
"Trained": 0.7787
},
{
"case": "medium_001",
"difficulty": "medium",
"Random": 0.233,
"Naive LLM": 0.196,
"Smart LLM": 0.3499,
"Trained": 0.2
},
{
"case": "hard_001",
"difficulty": "hard",
"Random": 0.2678,
"Naive LLM": 0.01,
"Smart LLM": 0.4839,
"Trained": 0.085
}
],
"averages": {
"case": "**Average**",
"difficulty": "",
"Random": 0.241,
"Naive LLM": 0.072,
"Smart LLM": 0.542,
"Trained": 0.3546
}
}