fix: audit sham projection before terminal collapse

1 files changed, 42 insertions, 9 deletions

diff --git a/experiments/analyze_kp_teaching_signal.py b/experiments/analyze_kp_teaching_signal.py
index db8649a..b850a81 100644
--- a/experiments/analyze_kp_teaching_signal.py
+++ b/experiments/analyze_kp_teaching_signal.py
@@ -127,12 +127,22 @@ def validate_record(path, rule, source_commit, bp_reference_macs):
             and max(value["maximum_observations"]
                     for value in projections) <= 128,
             f"{path}: projection batch observations")
-    require(max(value["maximum_post_projection_traffic_rms_ratio"]
-                for value in projections) <= 1e-5,
-            f"{path}: projected traffic remainder")
-    require(max(value["maximum_absolute_post_projection_soma_slope"]
-                for value in projections) <= 1e-5,
-            f"{path}: projected soma slope")
+    # Once a sham control has become nonfinite, numerical orthogonality of a
+    # projection that is not applied to its update is undefined. The frozen
+    # protocol explicitly retains such terminal collapses. Observation count
+    # and instruction leakage remain auditable for every batch; numerical
+    # projection quality is required only for a finite trajectory. Pretask
+    # predictor quality remains required for both cases.
+    require(finite(warmup["post_warmup_traffic_residual_rms_ratio"])
+            and warmup["post_warmup_traffic_residual_rms_ratio"] <= 1e-5,
+            f"{path}: pretask predictor residual")
+    if row["final"]["finite"]:
+        require(max(value["maximum_post_projection_traffic_rms_ratio"]
+                    for value in projections) <= 1e-5,
+                f"{path}: projected traffic remainder")
+        require(max(value["maximum_absolute_post_projection_soma_slope"]
+                    for value in projections) <= 1e-5,
+                f"{path}: projected soma slope")
 
     diagnostics = row["diagnostics"]
     if rule == "raw":
@@ -188,7 +198,7 @@ def main():
         ["git", "status", "--porcelain", "--untracked-files=no"],
         cwd=ROOT, check=True, capture_output=True, text=True).stdout.strip()
     require(not tracked_dirty, "analysis requires a clean tracked source")
-    source_commit = subprocess.run(
+    analysis_commit = subprocess.run(
         ["git", "rev-parse", "HEAD"], cwd=ROOT, check=True,
         capture_output=True, text=True).stdout.strip()
     with open(args.prerequisite) as handle:
@@ -203,9 +213,31 @@ def main():
                   prerequisite["metrics"]["accuracy"]),
             "dynamic record/gate mismatch")
 
+    record_paths = {
+        rule: os.path.join(args.outdir, f"{rule}.json") for rule in RULES
+    }
+    record_commits = set()
+    for path in record_paths.values():
+        with open(path) as handle:
+            record_commits.add(json.load(handle)["provenance"]["git_commit"])
+    require(len(record_commits) == 1,
+            "raw and matched controls must share one training source")
+    training_commit = record_commits.pop()
+    critical_training_paths = [
+        "KP_TEACHING_SIGNAL_ABLATION.md",
+        "experiments/conv_local_smoke.py",
+        "experiments/conv_run.py",
+        "experiments/kp_teaching_signal_ablation.py",
+        "sdil/conv.py",
+    ]
+    require(subprocess.run(
+        ["git", "diff", "--quiet", training_commit, analysis_commit, "--",
+         *critical_training_paths],
+        cwd=ROOT).returncode == 0,
+        "analysis revision changes a frozen training-critical path")
     rows = {
         rule: validate_record(
-            os.path.join(args.outdir, f"{rule}.json"), rule, source_commit,
+            record_paths[rule], rule, training_commit,
             prerequisite["metrics"]["bp_total_macs"])
         for rule in RULES
     }
@@ -246,7 +278,8 @@ def main():
             "control_mean_wall_s": statistics.mean(
                 row["timing"]["total_timed_wall_s"]
                 for row in rows.values()),
-            "source_commit": source_commit,
+            "training_source_commit": training_commit,
+            "analysis_source_commit": analysis_commit,
             "prerequisite_sha256": sha256(args.prerequisite),
         },
         "claim": (