improving-self-evolution-mbpp / round5_comparisons.json
humanlong's picture
Add files using upload-large-folder tool
0508458 verified
Raw History Blame Contribute Delete
764 Bytes
[
{
"candidate": "spectral_soft/round_5/evaluation",
"reference": "base/evaluation",
"pass1_delta": 0.0225,
"pass1_ci95": [
0.012,
0.0335
],
"noninferior_at_1pct": true,
"coverage4_delta": -0.3152965855006672,
"coverage4_ci95": [
-0.39445073313153484,
-0.2395721128340615
],
"paired_eligible_tasks": 245
},
{
"candidate": "spectral_soft/round_5/evaluation",
"reference": "spd_hard/round_5/evaluation",
"pass1_delta": -0.01175,
"pass1_ci95": [
-0.019,
-0.00475
],
"noninferior_at_1pct": false,
"coverage4_delta": 0.2929389999111115,
"coverage4_ci95": [
0.23322470344098753,
0.3531874808062307
],
"paired_eligible_tasks": 251
}
]