Auto-sync: 2026-06-29 07:26:06 (part 2)

results/paper_core_results.md

@@ -50,6 +50,7 @@ no-op prior is `+5.68 pp` over canonical h=16, same-state no-expert lattice is
 | K4 mean-by-type residual retrieval + wide scale-grid no-op prior 0.03 | No | No | 35.13% | +5.39 pp | Including 0.55 over-extends the transported tangent and drops below the best local calibration |
 | K4 mean-by-type residual retrieval + minimum-energy action penalty | No | No | 35.36-35.42% | +5.62-5.68 pp | A tiny action L2 penalty (0.05) ties the best row, while 0.10/0.20 drop slightly; shortest-action regularization does not add the gain |
 | K4 mean-by-type residual retrieval + source-advantage prior/gate | No | No | 35.13-35.30% | +5.39-5.57 pp | Measuring local train-source utility lift over the anchor does not replace the typed no-op prior; positive-advantage gates over-filter useful residual geometry |
+| K4 mean-by-type residual retrieval + train-family success bonus | No | No | 35.25-35.42% | +5.51-5.68 pp | A continuous train terminal-success prior is below the best by itself and only ties when added to the no-op row; train outcome reliability does not add the gain |
 | K4 mean-by-type residual retrieval + source-score prior 0.025 | No | No | 35.19% | +5.45 pp | A stronger reward-score prior drops below the plateau |
 | K4 mean-by-type residual retrieval, no-op-only residuals | No | No | 35.19% | +5.45 pp | Removing wrong-gripper residuals loses one success versus the fixed-scale safe-family plateau; the core gain is sparse no-op/tangent repair, with wrong-gripper acting only as a marginal helper |
 | K4 mean-by-type residual retrieval + margin sweep around 0.20 | No | No | 34.84-35.25% | +5.10-5.51 pp | Margin 0.20 is a local abstention optimum for both typed no-op and source-score priors; 0.15 and 0.25 drop below the plateau |
@@ -98,21 +99,22 @@ Suggested main-table rows:
 15. K4 mean-by-type residual retrieval + upper/wide tangent-length diagnostics
 16. K4 mean-by-type residual retrieval + minimum-energy action penalty diagnostics
 17. K4 mean-by-type residual retrieval + source-progress/source-score/source-advantage prior diagnostics
-18. K4 mean-by-type residual retrieval + no-op-only family diagnostic
-19. K4 mean-by-type residual retrieval + abstention margin fine sweep
-20. Source-progress viability gate diagnostics
-21. K2/K4 task-relative retrieval metric diagnostics
-22. K4 kernel-weighted residual consensus + no-op prior diagnostics
-23. K4 field-softmax residual barycenter + margin diagnostics
-24. K4 mean-by-type residual retrieval + wrong-gripper typed-prior diagnostics
-25. K2 broad tangent ray-search
-26. Residual-tangent distillation policy
-27. Residual+Gaussian hybrid, K32 sigma0.35
-28. Lattice, near-miss only
-29. Lattice, no expert
-30. Lattice, no expert + policy baseline candidate
-31. Lattice, full
-32. Oracle ceiling
+18. K4 mean-by-type residual retrieval + train-family success bonus diagnostics
+19. K4 mean-by-type residual retrieval + no-op-only family diagnostic
+20. K4 mean-by-type residual retrieval + abstention margin fine sweep
+21. Source-progress viability gate diagnostics
+22. K2/K4 task-relative retrieval metric diagnostics
+23. K4 kernel-weighted residual consensus + no-op prior diagnostics
+24. K4 field-softmax residual barycenter + margin diagnostics
+25. K4 mean-by-type residual retrieval + wrong-gripper typed-prior diagnostics
+26. K2 broad tangent ray-search
+27. Residual-tangent distillation policy
+28. Residual+Gaussian hybrid, K32 sigma0.35
+29. Lattice, near-miss only
+30. Lattice, no expert
+31. Lattice, no expert + policy baseline candidate
+32. Lattice, full
+33. Oracle ceiling
 
 Suggested claim:
 
@@ -125,7 +127,8 @@ Suggested claim:
 > simply shorter steps. Train-source progress/reward-score priors provide cleaner
 > fixed-scale ties but not the top row; source-advantage priors/gates are negative,
 > suggesting transferable residuals need not beat the expert anchor in their source
-> state. Ungated KNN residual
+> state. Continuous train-family success priors likewise tie or drop rather than
+> explain the top row. Ungated KNN residual
 > retrieval, field-gradient ascent, broader non-expert BC targets, field-teacher/tangent distillation, z-score/task-relative retrieval,
 > train-family reliability priors, policy-relative anchoring, residual+Gaussian hybrids,
 > source-progress/source-advantage viability gates, no-op-only family masking, off-peak abstention margins, overly strong train-outcome priors, tangent consensus, kernel-weighted tangent interpolation, field-softmax tangent barycenters, tangent ray-search, wrong-gripper typed priors, and same-state policy-baseline fallback fail to improve the main rows.

results/paper_story_memo.md

@@ -31,6 +31,7 @@ when queried on proposal geometry that matches those local counterfactuals.
 | Field-gated tangent length calibration improves the clean bridge | K4 mean-by-type scale grid 0.35/0.40/0.45 with no-op bonus 0.03 reaches 35.42%; the source-score version reaches 35.30%. Upper 0.40/0.45/0.50 nearly ties at 35.36%, while wide 0.35/0.45/0.55 drops to 35.13% | Current best clean result; local scale calibration, not a larger-step effect |
 | Minimum-energy residual regularization does not add the gain | action L2 penalty 0.05 ties 35.42%, while 0.10/0.20 reach 35.36% | Negative/tie diagnostic: the clean bridge is not explained by shortest-action bias |
 | Source-advantage priors/gates are too brittle | source-advantage bonuses 0.02/0.05 reach 35.13%; no-op+advantage bonus reaches 35.30%; positive-advantage gates reach 35.13% with or without no-op prior | Negative diagnostic: useful transferable tangents need not beat the expert anchor in their own source state |
+| Continuous train-family success priors do not add the gain | scale-grid family-success bonuses 0.02/0.03/0.05 reach 35.25%; no-op+family-success 0.02 ties the best at 35.42% | Negative/tie diagnostic: train terminal success is not the right confidence signal for transferred tangents |
 | Kernel-weighted tangent interpolation does not beat equal consensus | K4 kernel-weighted residual consensus reaches 34.96%; with no-op prior and scales 0.35/0.40/0.45 it reaches 35.13%/35.19%/35.19%, below the 35.25% mean-consensus plateau | Negative/near-tie diagnostic |
 | Field-conditioned tangent barycenters identify good sparse corrections but do not close the proposal gap | K4 field-softmax transport reaches 34.96%; with no-op prior and margins 0.10/0.05/0.00 it reaches 35.19%/35.07%/34.84%. Selected aggregate residuals are high-value (up to 60.00% success), but selecting more of them degrades the global row | Negative/near-tie diagnostic |
 | Tangent ray-search does not beat the typed-prior clean row | K1/K2 tight scale-grid ray search reach 34.84%; K2 broad reaches 34.96%; K4 tight reaches 34.55%, all below the scale-grid mean-consensus row at 35.42% | Near-tie/negative diagnostic |
@@ -76,25 +77,26 @@ clean proposal result, the intended main rows are:
 19. K4 mean-by-type tangent consensus + train-source progress prior: 35.25% at bonus 0.03; 35.13% at bonus 0.05
 20. K4 mean-by-type tangent consensus + train-source reward-score prior: 35.25% at bonuses 0.015/0.020; 35.30% with scale grid; 35.19% at 0.025
 21. K4 mean-by-type tangent consensus + train-source advantage prior/gate: 35.13% at bonuses 0.02/0.05; 35.30% with no-op+advantage; 35.13% with positive-advantage gates
-22. K4 mean-by-type tangent consensus, no-op-only residuals: 35.19% with either no-op bonus 0.03 or source-score bonus 0.02
-23. K4 mean-by-type abstention margin sweep: 35.07% / 35.25% / 34.84% for typed no-op margins 0.15 / 0.20 / 0.25; 34.96% / 35.25% / 34.84% for source-score margins
-24. Source-progress viability gates: 35.19% / 34.96% / 34.72% for thresholds 0.25 / 0.50 / 0.75
-25. K4 kernel-weighted tangent consensus / + no-op prior: 34.96% / 35.19%
-26. K4 field-softmax tangent transport / best margin sweep: 34.96% / 35.19%
-27. Wrong-gripper prior / no-op+wrong-gripper prior: 35.19% / 35.25%
-27. K2 broad tangent ray-search: 34.96%
-28. K1/K2 tight tangent ray-search: 34.84% / 34.84%
-29. K4 tight tangent ray-search: 34.55%
-30. Residual-tangent distillation policy: 28.87%
-31. Z-score residual retrieval: 32.23-32.81%
-32. Task-relative residual retrieval metric: 34.26-34.43%
-33. Train-family reliability prior: 33.28-33.33%
-34. Residual+Gaussian hybrid K32/K64: 31.30% / 30.90%
-35. Lattice, near-miss only: 55.94%
-36. Lattice, no expert: 56.99%
-37. Lattice, no expert + policy baseline candidate: 40.70%
-38. Lattice, full: 69.33%
-39. Oracle ceiling: 86.78%
+22. K4 mean-by-type tangent consensus + train-family success bonus: 35.25% alone; 35.42% with no-op bonus 0.03
+23. K4 mean-by-type tangent consensus, no-op-only residuals: 35.19% with either no-op bonus 0.03 or source-score bonus 0.02
+24. K4 mean-by-type abstention margin sweep: 35.07% / 35.25% / 34.84% for typed no-op margins 0.15 / 0.20 / 0.25; 34.96% / 35.25% / 34.84% for source-score margins
+25. Source-progress viability gates: 35.19% / 34.96% / 34.72% for thresholds 0.25 / 0.50 / 0.75
+26. K4 kernel-weighted tangent consensus / + no-op prior: 34.96% / 35.19%
+27. K4 field-softmax tangent transport / best margin sweep: 34.96% / 35.19%
+28. Wrong-gripper prior / no-op+wrong-gripper prior: 35.19% / 35.25%
+29. K2 broad tangent ray-search: 34.96%
+30. K1/K2 tight tangent ray-search: 34.84% / 34.84%
+31. K4 tight tangent ray-search: 34.55%
+32. Residual-tangent distillation policy: 28.87%
+33. Z-score residual retrieval: 32.23-32.81%
+34. Task-relative residual retrieval metric: 34.26-34.43%
+35. Train-family reliability prior: 33.28-33.33%
+36. Residual+Gaussian hybrid K32/K64: 31.30% / 30.90%
+37. Lattice, near-miss only: 55.94%
+38. Lattice, no expert: 56.99%
+39. Lattice, no expert + policy baseline candidate: 40.70%
+40. Lattice, full: 69.33%
+41. Oracle ceiling: 86.78%
 
 ## Novelty Framing
 
@@ -122,10 +124,12 @@ test-time search. The cleaner novelty is:
 
 ## Job Status
 
-Last checked: `2026-06-29 10:57 UTC`. The K4 mean-by-type scale-grid sweep
+Last checked: `2026-06-29 11:25 UTC`. The K4 mean-by-type scale-grid sweep
 completed and produced a new clean best, 35.42%, while upper/wide,
-minimum-energy, and source-advantage follow-ups completed without improving it.
-The paper table/paired analysis use that row as `best_clean_key`.
+minimum-energy, source-advantage, and train-family success-prior follow-ups
+completed without improving it. Consensus-confidence jobs are running/pending
+as the next geometric reliability test. The paper table/paired analysis use
+the scale-grid no-op row as `best_clean_key` unless those jobs improve it.
 
 - `14858328`-`14858333`: completed train-split `field_selected_noexpert_bc5`;
   direct rollout is 26.84%, field-guided best is 27.65%.
@@ -238,6 +242,17 @@ The paper table/paired analysis use that row as `best_clean_key`.
   Summary jobs `14893788`/`14893790` and rebuild job `14893791` completed. This
   suggests that raw full-state similarity still carries useful robot/phase
   information for residual transfer; object-only actor pose is too lossy here.
+- `14903128`/`14903130`/`14903132`/`14903134`: completed continuous
+  train-family success-prior GPU arrays. Family-success bonuses `0.02`, `0.03`,
+  and `0.05` reach 35.25%; adding family-success `0.02` to the no-op `0.03`
+  best row ties 35.42% without adding a new gain. Summary jobs `14903129`/
+  `14903131`/`14903133`/`14903135` and rebuild job `14903136` completed.
+- `14903296`: completed CPU smoke for the train-neighbor consensus-confidence
+  penalty path, validating metadata and Slurm/CLI wiring.
+- `14903384`/`14903386`/`14903388`/`14903390`: submitted consensus-confidence
+  GPU arrays for consensus-only `0.05` and no-op `0.03` plus consensus penalties
+  `0.02`, `0.05`, and `0.10`. Summary jobs are `14903385`/`14903387`/
+  `14903389`/`14903391`; rebuild job `14903392` depends on those summaries.
 - `14894281`: completed the Apptainer unit smoke for the train-source
   progress-viability gate, including the variable residual-count padding check
   (`source_progress_lengths == [3, 3]`).

scripts/build_paper_analysis.py

@@ -283,6 +283,38 @@ METHODS = [
             "k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_typesuccessbonus0p02_summary.json"
         ),
     ),
+    MethodSpec(
+        key="residual_k4_consensus_grid035040045_consensus005",
+        label="K4 mean-by-type tangent consensus, scales 0.35/0.40/0.45, consensus penalty 0.05",
+        summary_path=(
+            "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_"
+            "k4_grid035040045_safe_margin0p20_mean_by_type_consensus0p05_summary.json"
+        ),
+    ),
+    MethodSpec(
+        key="residual_k4_consensus_grid035040045_noopbonus003_consensus002",
+        label="K4 mean-by-type tangent consensus, scales 0.35/0.40/0.45, no-op bonus 0.03, consensus penalty 0.02",
+        summary_path=(
+            "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_"
+            "k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_consensus0p02_summary.json"
+        ),
+    ),
+    MethodSpec(
+        key="residual_k4_consensus_grid035040045_noopbonus003_consensus005",
+        label="K4 mean-by-type tangent consensus, scales 0.35/0.40/0.45, no-op bonus 0.03, consensus penalty 0.05",
+        summary_path=(
+            "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_"
+            "k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_consensus0p05_summary.json"
+        ),
+    ),
+    MethodSpec(
+        key="residual_k4_consensus_grid035040045_noopbonus003_consensus010",
+        label="K4 mean-by-type tangent consensus, scales 0.35/0.40/0.45, no-op bonus 0.03, consensus penalty 0.10",
+        summary_path=(
+            "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_"
+            "k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_consensus0p10_summary.json"
+        ),
+    ),
     MethodSpec(
         key="residual_k4_consensus_grid035040045_noopbonus003_l2penalty005",
         label="K4 mean-by-type tangent consensus, scales 0.35/0.40/0.45, no-op bonus 0.03, action L2 penalty 0.05",

scripts/build_paper_table_status.py

@@ -615,6 +615,46 @@ SPECS = [
         story_role="continuous train-family reliability calibration on the current best typed prior",
         pending_job="14903134/14903135",
     ),
+    ResultSpec(
+        key="retrieval_residual_k4_mean_grid035040045_consensus005",
+        label="K4 mean-by-type residual retrieval, scales 0.35/0.40/0.45, margin 0.20, consensus penalty 0.05",
+        path="h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_grid035040045_safe_margin0p20_mean_by_type_consensus0p05_summary.json",
+        clean_deployment="yes",
+        same_state_proposals="no",
+        expert_proposal="no",
+        story_role="train-neighbor tangent-consensus confidence without sparse type prior",
+        pending_job="14903384/14903385",
+    ),
+    ResultSpec(
+        key="retrieval_residual_k4_mean_grid035040045_noopbonus003_consensus002",
+        label="K4 mean-by-type residual retrieval, scales 0.35/0.40/0.45, margin 0.20, no-op bonus 0.03, consensus penalty 0.02",
+        path="h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_consensus0p02_summary.json",
+        clean_deployment="yes",
+        same_state_proposals="no",
+        expert_proposal="no",
+        story_role="train-neighbor tangent-consensus confidence on the current best typed prior",
+        pending_job="14903386/14903387",
+    ),
+    ResultSpec(
+        key="retrieval_residual_k4_mean_grid035040045_noopbonus003_consensus005",
+        label="K4 mean-by-type residual retrieval, scales 0.35/0.40/0.45, margin 0.20, no-op bonus 0.03, consensus penalty 0.05",
+        path="h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_consensus0p05_summary.json",
+        clean_deployment="yes",
+        same_state_proposals="no",
+        expert_proposal="no",
+        story_role="train-neighbor tangent-consensus confidence on the current best typed prior",
+        pending_job="14903388/14903389",
+    ),
+    ResultSpec(
+        key="retrieval_residual_k4_mean_grid035040045_noopbonus003_consensus010",
+        label="K4 mean-by-type residual retrieval, scales 0.35/0.40/0.45, margin 0.20, no-op bonus 0.03, consensus penalty 0.10",
+        path="h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_grid035040045_safe_margin0p20_mean_by_type_noopbonus0p03_consensus0p10_summary.json",
+        clean_deployment="yes",
+        same_state_proposals="no",
+        expert_proposal="no",
+        story_role="train-neighbor tangent-consensus confidence on the current best typed prior",
+        pending_job="14903390/14903391",
+    ),
     ResultSpec(
         key="retrieval_residual_k4_mean_grid035040045_noopbonus003_l2penalty005",
         label="K4 mean-by-type residual retrieval, scales 0.35/0.40/0.45, margin 0.20, no-op bonus 0.03, action L2 penalty 0.05",