Kusshal commited on
Commit
f0c5d7d
·
verified ·
1 Parent(s): b0fbfab

sync EmbedEd artifacts (phase01-v6)

Browse files
artifacts/gate_runs.json CHANGED
@@ -96,6 +96,38 @@
96
  }
97
  }
98
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
99
  {
100
  "stamp": null,
101
  "version": "phase01-v6",
 
96
  }
97
  }
98
  },
99
+ {
100
+ "stamp": null,
101
+ "version": "phase01-v6",
102
+ "rule": "D1 scale-aware: C1 < C2 <= C3 and d(C1->C2) >= 0.5",
103
+ "ok": true,
104
+ "d_c1_c2": 0.614,
105
+ "ci": [
106
+ 0.795,
107
+ 1.202
108
+ ],
109
+ "recorded_source": null,
110
+ "per_condition": {
111
+ "C1": {
112
+ "mean_cos": 0.96068,
113
+ "std": 0.0257,
114
+ "n_samples": 20000,
115
+ "n_anchors": 1000
116
+ },
117
+ "C2": {
118
+ "mean_cos": 0.97526,
119
+ "std": 0.02162,
120
+ "n_samples": 20000,
121
+ "n_anchors": 1000
122
+ },
123
+ "C3": {
124
+ "mean_cos": 0.98845,
125
+ "std": 0.00722,
126
+ "n_samples": 20000,
127
+ "n_anchors": 1000
128
+ }
129
+ }
130
+ },
131
  {
132
  "stamp": null,
133
  "version": "phase01-v6",
report/measurements.md CHANGED
@@ -525,6 +525,17 @@ Context recorded at the time: ordering C1 < C2 ≤ C3 holds; the C1→C2 gap is
525
  - legacy absolute-margin rule (0.02): FAIL — C1→C2 gap +0.0146 < margin 0.02 (mining produced equally easy negatives) (reported for the record; not the verdict since v6)
526
  - **verdict: PASS** — order C1 < C2 <= C3 holds (0.9607 < 0.9753 <= 0.9885) and d(C1→C2) = 0.614 >= 0.5 (absolute gap +0.0146)
527
 
 
 
 
 
 
 
 
 
 
 
 
528
  ## Gate G1 — diagnostics (scale of the space; no verdict)
529
 
530
  `python -m scripts.hardness_diagnostics` — version `phase01-v6`, 1600 anchors × k=20, corpus 6451. Percentile = share of corpus fragments less similar to the anchor than the negative (50 = random, 100 = nearest).
 
525
  - legacy absolute-margin rule (0.02): FAIL — C1→C2 gap +0.0146 < margin 0.02 (mining produced equally easy negatives) (reported for the record; not the verdict since v6)
526
  - **verdict: PASS** — order C1 < C2 <= C3 holds (0.9607 < 0.9753 <= 0.9885) and d(C1→C2) = 0.614 >= 0.5 (absolute gap +0.0146)
527
 
528
+ ### run 2026-09-29 18:18 UTC — version `phase01-v6`, rule: D1 scale-aware: C1 < C2 <= C3 and d(C1->C2) >= 0.5
529
+
530
+ - C1: mean cos(anchor, negative) = 0.96068 (sd 0.026, n = 20000, anchors = 1000)
531
+ - C2: mean cos(anchor, negative) = 0.97526 (sd 0.022, n = 20000, anchors = 1000)
532
+ - C3: mean cos(anchor, negative) = 0.98845 (sd 0.007, n = 20000, anchors = 1000)
533
+ - corpus percentile of the negative in its anchor's similarity ranking, median: C1 50.1 / C2 91.9 / C3 99.8 (50 = random, 100 = nearest); inside the anchor's top-20: C1 0.3% / C2 15.4% / C3 94.8%
534
+ - embeddings: microsoft/graphcodebert-base max_len=512 n=8063 version=phase01-v5
535
+ - d(C1→C2) = 0.614, 95% CI 0.795–1.202 (2000 resamples over anchors, not negatives)
536
+ - legacy absolute-margin rule (0.02): FAIL — C1→C2 gap +0.0146 < margin 0.02 (mining produced equally easy negatives) (reported for the record; not the verdict since v6)
537
+ - **verdict: PASS** — order C1 < C2 <= C3 holds (0.9607 < 0.9753 <= 0.9885) and d(C1→C2) = 0.614 >= 0.5 (absolute gap +0.0146)
538
+
539
  ## Gate G1 — diagnostics (scale of the space; no verdict)
540
 
541
  `python -m scripts.hardness_diagnostics` — version `phase01-v6`, 1600 anchors × k=20, corpus 6451. Percentile = share of corpus fragments less similar to the anchor than the negative (50 = random, 100 = nearest).
sync_manifest.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "version": "phase01-v6",
3
- "synced_at_utc": "2026-09-29T17:56:21Z",
4
  "staged_files": [
5
  "artifacts/audit/audit_key.csv",
6
  "artifacts/audit/audit_labels.csv",
@@ -90,7 +90,8 @@
90
  "artifacts/v5-backup/mining_summary.json",
91
  "artifacts/v5-backup/triples_C1.jsonl",
92
  "artifacts/v5-backup/triples_C2.jsonl",
93
- "artifacts/v5-backup/triples_C3.jsonl"
 
94
  ],
95
  "excluded_patterns": [
96
  "sprime/*",
 
1
  {
2
  "version": "phase01-v6",
3
+ "synced_at_utc": "2026-09-29T18:18:32Z",
4
  "staged_files": [
5
  "artifacts/audit/audit_key.csv",
6
  "artifacts/audit/audit_labels.csv",
 
90
  "artifacts/v5-backup/mining_summary.json",
91
  "artifacts/v5-backup/triples_C1.jsonl",
92
  "artifacts/v5-backup/triples_C2.jsonl",
93
+ "artifacts/v5-backup/triples_C3.jsonl",
94
+ "report/measurements.md"
95
  ],
96
  "excluded_patterns": [
97
  "sprime/*",