Ektome-Qwen2.5-Coder-14B-Instruct-PristinelyUncensored / cert_Qwen2.5-Coder-14B-Instruct.json
Zynerji's picture
Sphragis certificate n=2800: PASS (capability-aware re-abliteration)
8bceebc verified
Raw
History Blame
3.22 kB
{
"sphragis_version": "0.1.0",
"generated_at": "2026-07-27T03:53:12.900485+00:00",
"reference": {
"endpoint": "http://127.0.0.1:8080/v1",
"model": "ref"
},
"candidate": {
"endpoint": "http://127.0.0.1:8081/v1",
"model": "cand"
},
"pack": {
"name": "/root/pack_v2.jsonl",
"n_tasks": 2800,
"sha256": "2de27099bbb15bab4f7b35599b215f038fdb68b3f5f4e1cda1a464f9dd18e14e"
},
"overall": "PASS",
"claim": "For every axis, the candidate was demonstrated non-inferior to the reference: the one-sided 95% upper confidence bound on the accuracy regression is below the margin of 3.0%.",
"axes": [
{
"axis": "arithmetic",
"n": 1400,
"counts": {
"both_correct": 1301,
"ref_only": 4,
"cand_only": 1,
"both_wrong": 94
},
"acc_reference": 0.932143,
"acc_candidate": 0.93,
"regression_d": 0.002143,
"d_ci": [
-0.000714,
0.005714
],
"d_upper_bound": 0.005,
"p_regression": 0.1875,
"p_regression_holm": 0.75,
"p_improvement": 0.96875,
"improved": false,
"mde_at_power": 0.003571,
"n_needed_for_margin": 204,
"verdict": "PASS",
"reason": "non_inferior_within_margin"
},
{
"axis": "instruction",
"n": 600,
"counts": {
"both_correct": 454,
"ref_only": 0,
"cand_only": 0,
"both_wrong": 146
},
"acc_reference": 0.756667,
"acc_candidate": 0.756667,
"regression_d": 0.0,
"d_ci": [
0.0,
0.0
],
"d_upper_bound": 0.0,
"p_regression": 1.0,
"p_regression_holm": 1.0,
"p_improvement": 1.0,
"improved": false,
"mde_at_power": null,
"n_needed_for_margin": 204,
"verdict": "PASS",
"reason": "non_inferior_within_margin"
},
{
"axis": "knowledge",
"n": 400,
"counts": {
"both_correct": 380,
"ref_only": 7,
"cand_only": 9,
"both_wrong": 4
},
"acc_reference": 0.9675,
"acc_candidate": 0.9725,
"regression_d": -0.005,
"d_ci": [
-0.025,
0.015
],
"d_upper_bound": 0.01,
"p_regression": 0.77275085,
"p_regression_holm": 1.0,
"p_improvement": 0.40180969,
"improved": false,
"mde_at_power": 0.0248,
"n_needed_for_margin": 273,
"verdict": "PASS",
"reason": "non_inferior_within_margin"
},
{
"axis": "reasoning",
"n": 400,
"counts": {
"both_correct": 371,
"ref_only": 0,
"cand_only": 2,
"both_wrong": 27
},
"acc_reference": 0.9275,
"acc_candidate": 0.9325,
"regression_d": -0.005,
"d_ci": [
-0.0125,
0.0
],
"d_upper_bound": 0.0,
"p_regression": 1.0,
"p_regression_holm": 1.0,
"p_improvement": 0.25,
"improved": false,
"mde_at_power": 0.005,
"n_needed_for_margin": 204,
"verdict": "PASS",
"reason": "non_inferior_within_margin"
}
],
"params": {
"margin": 0.03,
"alpha": 0.05,
"n_floor": 30,
"power": 0.8,
"n_boot": 4000,
"seed": 0
}
}