Auto-sync: 2026-06-29 02:18:16

results/README.md

@@ -101,9 +101,9 @@ Completed deployment-clean diagnostics:
 
 Completed residual transport jobs:
 
-Last checked: `2026-06-29 05:54 UTC`. The source-score bonus-prior batch
-completed after passing CPU/unit smokes, and the paper table/paired analysis now
-include progress and reward-score train-outcome priors.
+Last checked: `2026-06-29 06:13 UTC`. The no-op-only residual-family ablation
+completed after CPU smokes and GPU arrays, and the paper table/paired analysis
+now include both full safe-family and no-op-only train-outcome prior rows.
 
 - Code supports `--retrieval-residual-scale`, defaulting to `1.0` so the
   completed 32.12% result is unchanged.
@@ -226,6 +226,15 @@ include progress and reward-score train-outcome priors.
   Bonuses `0.015` and `0.020` tie the current best at 35.25%, while bonus
   `0.025` reaches 35.19%. Summary jobs `14897126`/`14897127`/`14897128` and
   rebuild job `14897129` completed.
+- Jobs `14897548`/`14897549` completed no-op-only CPU rollout smokes after
+  excluding wrong-gripper residuals from the safe family.
+- Jobs `14897563`/`14897564` completed no-op-only GPU arrays. Both no-op bonus
+  `0.03` and source-score bonus `0.02` reach 35.19%, one success below the
+  35.25% best safe-family plateau. This shows wrong-gripper residuals are not
+  the main mechanism but still provide a small marginal tie-breaking gain.
+  Summary jobs `14897565`/`14897566` and rebuild job `14897567` were submitted;
+  the same summaries/analysis were also rebuilt locally while the CPU summary
+  job was pending.
 - Job `14869627` completed the CPU Apptainer smoke for the residual scale-grid
   selector: selected index `3` and returned action `0.20` on the toy ray-search
   case, validating candidate expansion before rollout.