sync EmbedEd artifacts (phase01-v6)

.gitattributes

@@ -49,3 +49,4 @@ artifacts/sprime/Clones[[:space:]]paper[[:space:]]replication[[:space:]]package[
 artifacts/sprime/Clones[[:space:]]paper[[:space:]]replication[[:space:]]package[[:space:]]release[[:space:]]zip/oj_dataset_experiments/astnn/data/c/oj_and_scb/pair_ids.pkl filter=lfs diff=lfs merge=lfs -text
 artifacts/sprime/Clones[[:space:]]paper[[:space:]]replication[[:space:]]package[[:space:]]release[[:space:]]zip/oj_dataset_experiments/astnn/data/c/oj_and_scb/programs.pkl filter=lfs diff=lfs merge=lfs -text
 artifacts/corpus_emb.npy filter=lfs diff=lfs merge=lfs -text
+artifacts/runs/C0_13/predictions.npz filter=lfs diff=lfs merge=lfs -text

artifacts/runs/C0_13/eval_config.json

@@ -0,0 +1,13 @@
+{
+  "condition": "C0",
+  "seed": 13,
+  "smoke": false,
+  "version": "phase01-v6",
+  "model_id": "microsoft/graphcodebert-base",
+  "pooling": "mean",
+  "max_len": 512,
+  "eval_batch": 64,
+  "threshold_policy": "max_f1_on_valid",
+  "pairs_source": "artifacts/pairs_{valid,test}.tsv",
+  "fingerprint": "670ec216db20b468"
+}

artifacts/runs/C0_13/eval_metrics.json

@@ -0,0 +1,37 @@
+{
+  "condition": "C0",
+  "seed": 13,
+  "version": "phase01-v6",
+  "fingerprint": "670ec216db20b468",
+  "smoke": false,
+  "model": "microsoft/graphcodebert-base (untuned)",
+  "threshold_policy": "max_f1_on_valid",
+  "threshold": 0.977894,
+  "valid": {
+    "precision": 0.2029,
+    "recall": 0.3378,
+    "f1": 0.2536,
+    "n": 415416
+  },
+  "valid_map_at_r": 0.1021,
+  "test_f1": 0.2572,
+  "test_precision": 0.2075,
+  "test_recall": 0.3383,
+  "test_map_at_r": 0.1024,
+  "test": {
+    "precision": 0.2075,
+    "recall": 0.3383,
+    "f1": 0.2572,
+    "tp": 19220,
+    "fp": 73401,
+    "fn": 37600,
+    "n": 415416
+  },
+  "n_fragments_encoded": 1685,
+  "n_pairs": {
+    "valid": 415416,
+    "test": 415416
+  },
+  "encode_seconds": 14.9,
+  "evaluated_utc": "2026-09-27T16:46:21Z"
+}

artifacts/runs/C0_13/predictions.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ed148d7a85dfd07d6006931474cf9c06452109e85c776b1387c41402420cf5f
+size 2800167

report/measurements.md

@@ -533,3 +533,18 @@ Label-noise floors (correction 5; proxies, not the audit): share of negatives th
 | C2 | 50 | 21 | 29 | 0 | **42%** [29%, 56%] | 42% [29%, 56%] |
 | C3 | 50 | 15 | 34 | 1 | **30%** [19%, 44%] | 32% [21%, 46%] |
 
+## Phase 2 — main results
+
+Generated from `artifacts/runs/<condition>_<seed>/eval_metrics.json` by `python -m embeded.evaluate --results-table` — version `phase01-v6`. Threshold policy: *max_f1_on_valid* (chosen on validation, applied unchanged to test; test is scored once per run).
+
+| condition | seed | F1 | precision | recall | MAP@R | valid threshold | n test pairs |
+|---|---|---|---|---|---|---|---|
+| C0 | 13 | 0.2572 | 0.2075 | 0.3383 | 0.1024 | 0.9779 | 415416 |
+
+| condition | runs | mean F1 | sd F1 | mean P | mean R | mean MAP@R |
+|---|---|---|---|---|---|---|
+| C0 | 1 | 0.2572 | 0.0000 | 0.2075 | 0.3383 | 0.1024 |
+
+- C0 is the untuned baseline (one evaluation, not a trained condition) at mean F1 0.2572.
+- Sanity reference: published CodeXGLUE fine-tuned-CodeBERT F1 ~ 0.95 under a different model/protocol. Observed F1 range 0.2572–0.2572. A large discrepancy is a harness/protocol check (metric implementation, pair order, labels, threshold handling), not a reason to tune against the test split.
+

sync_manifest.json

@@ -1,6 +1,6 @@
 {
   "version": "phase01-v6",
-  "synced_at_utc": "2026-09-27T16:45:33Z",
+  "synced_at_utc": "2026-09-27T17:03:35Z",
   "staged_files": [
     "artifacts/audit/audit_key.csv",
     "artifacts/audit/audit_labels.csv",
@@ -18,6 +18,9 @@
     "artifacts/pairs_test.tsv",
     "artifacts/pairs_train.tsv",
     "artifacts/pairs_valid.tsv",
+    "artifacts/runs/C0_13/eval_config.json",
+    "artifacts/runs/C0_13/eval_metrics.json",
+    "artifacts/runs/C0_13/predictions.npz",
     "artifacts/token_lengths.json",
     "artifacts/triples_C1.jsonl",
     "artifacts/triples_C2.jsonl",