Auto-sync: 2026-06-29 12:01:54 (part 2)

results/README.md

@@ -64,6 +64,8 @@ Use these summaries for the current paper tables:
 	  `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_grid035045055_safe_margin0p20_mean_by_type_noopbonus0p03_summary.md`,
 	  `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_fieldsoftmax_grid_safe_margin0p20_summary.md`,
 	  `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_fieldsoftmax_grid_safe_margin0p10_noopbonus0p03_summary.md`,
+	  `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_grid035040045_safe_margin0p20_summary.md`,
+	  `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_grid035040045_safe_margin0p20_noopbonus0p03_summary.md`,
   `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_policy_anchor_scale0p35_safe_noexpert_summary.md`,
   `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_scale0p35_type_success010_summary.md`,
   `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_scale0p35_type_success025_summary.md`
@@ -88,7 +90,7 @@ Known stale or incomplete files:
 
 Completed deployment-clean diagnostics:
 
-- Job `14837136` swept K/sigma for the current best deployment-clean route
+- Job `14837136` swept K/sigma for an early deployment-clean route
   (`near_miss_policy_bc5` + `best.pt` + Gaussian field selection). Summary job
   `14837137` wrote `h16_policy_ckpt_near_miss_policy_bc5_bestpt_field_sweep_summary.*`.
 - Job `14842528` completed a partial trust-region `field_optim` sweep; the
@@ -114,13 +116,13 @@ Completed deployment-clean diagnostics:
 
 Completed residual transport jobs:
 
-Last checked: `2026-06-29 12:03 UTC`. The K4 mean-by-type scale-grid sweep
-completed and produced a new clean best, 35.42%, while upper/wide,
-minimum-energy, source-advantage, and train-family success-prior follow-ups
-completed without improving it. Consensus-confidence follow-ups also completed
-below the best row. Repair-tangent follow-ups also completed below the best
-row, with a best repair result of 34.43%. The paper table/paired analysis now
-use the scale-grid no-op row as `best_clean_key`.
+Last checked: `2026-06-29 16:12 UTC`. The K4 masked composed type-consensus
+sweep completed and produced a new clean best, 35.54%, while the pure masked
+composition row reached 35.30%. Raw selected candidate types show no
+random-negative or wrong-direction composite leak. The paper table/paired
+analysis now use the masked composed no-op row as `best_clean_key`. A
+component-wise composite-prior follow-up has passed CPU smoke and is queued for
+3-seed GPU rollout.
 
 - Code supports `--retrieval-residual-scale`, defaulting to `1.0` so the
   completed 32.12% result is unchanged.
@@ -223,23 +225,36 @@ use the scale-grid no-op row as `best_clean_key`.
 - Jobs `14903128`/`14903130`/`14903132`/`14903134` completed the continuous
   train-family success-prior sweep. Family-success bonuses `0.02`, `0.03`, and
   `0.05` reach 35.25%; adding family-success `0.02` to no-op `0.03` ties the
-  35.42% scale-grid best without improving it. Summary jobs `14903129`/
+  previous 35.42% scale-grid best without improving it. Summary jobs `14903129`/
   `14903131`/`14903133`/`14903135` and rebuild job `14903136` completed.
 - Job `14903296` completed the CPU smoke for train-neighbor
   consensus-confidence penalties. Jobs `14903384`/`14903386`/`14903388`/
   `14903390` completed the GPU arrays for consensus-only `0.05` and no-op
   `0.03` plus consensus penalties `0.02`, `0.05`, and `0.10`. Results are
   35.19% for consensus-only and 35.36% for all no-op+consensus rows, below the
-  35.42% best. Summaries `14903385`/`14903387`/`14903389`/`14903391` and rebuild
+  previous 35.42% scale-grid best. Summaries `14903385`/`14903387`/`14903389`/`14903391` and rebuild
   `14903392` completed.
 - Job `14904575` completed the CPU smoke for repair-tangent residual direction
   (`anchor_minus_candidate`), validating the new metadata, CLI, and Slurm
   passthrough. Jobs `14904737`/`14904740`/`14904742`/`14904744` completed the
   repair-tangent GPU arrays. Near-miss-only repair grids reach 34.14-34.38%,
   and the safe-family repair row reaches 34.43%, so transported
-  failure-to-expert corrections are a negative diagnostic below the 35.42% best.
+  failure-to-expert corrections are a negative diagnostic below the previous
+  35.42% scale-grid best.
   Summary jobs `14904738`/`14904741`/`14904743`/`14904745` completed; local
   rebuilds updated paper artifacts, and rebuild job `14904803` was canceled.
+- Job `14911977` completed the CPU smoke for masked composed type-consensus
+  transport and confirmed composite masking excludes random-negative and
+  wrong-direction parts. Jobs `14911979`/`14911980` completed the K4 masked
+  composition GPU arrays. Pure masked composition reaches 35.30%; adding the
+  typed no-op prior reaches the new clean best, 35.54%. Summary jobs `14911982`/
+  `14911983` and rebuild job `14911984` completed.
+- Job `14912552` completed the CPU smoke for component-wise candidate-type
+  bonuses on masked composed type-consensus transport. Metadata records
+  `candidate_type_bonus_components=True`, selected candidate types have no
+  random-negative or wrong-direction leak, and the smoke selected
+  `policy_residual` on 8/8 groups. Jobs `14912561`/`14912562`/`14912563` are
+  queued for the 3-seed GPU array, summary, and paper-artifact rebuild.
 - Job `14894281` completed the unit smoke for the source-progress viability gate,
   including the variable residual-count padding check.
 - Job `14894282` completed the ManiSkill CPU rollout smoke for K4 mean-by-type
@@ -254,13 +269,13 @@ use the scale-grid no-op row as `best_clean_key`.
 - Jobs `14894672`/`14894673` completed unit and CPU rollout smokes for the
   continuous train-source progress bonus path.
 - Jobs `14894674`/`14894675` completed source-progress bonus arrays with no fixed
-  no-op prior. Bonus `0.03` ties the current best at 35.25%, while bonus `0.05`
+  no-op prior. Bonus `0.03` ties the fixed-scale plateau at 35.25%, while bonus `0.05`
   reaches 35.13%. Summary jobs `14894676`/`14894677` completed; rebuild job
   `14894678` was queued after them.
 - Jobs `14897121`/`14897122` completed unit and CPU rollout smokes for the
   train-source reward-score bonus path.
 - Jobs `14897123`/`14897124`/`14897125` completed source-score bonus arrays.
-  Bonuses `0.015` and `0.020` tie the current best at 35.25%, while bonus
+  Bonuses `0.015` and `0.020` tie the fixed-scale plateau 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
@@ -279,15 +294,15 @@ use the scale-grid no-op row as `best_clean_key`.
   `14897845`-`14897848` and rebuild job `14897849` completed.
 - Jobs `14897988`/`14897989` completed K4 mean-by-type scale-grid sweeps using
   scales `0.35/0.40/0.45`, margin `0.20`, and safe residual families. The
-  typed no-op prior row reaches a new clean best, 35.42%; the source-score prior
+  typed no-op prior row reaches a then-new clean best, 35.42%; the source-score prior
   row reaches 35.30%. Summary jobs `14897990`/`14897991` completed; rebuild job
   `14897992` was submitted, and local rebuilds updated the paper artifacts.
 - Jobs `14898107`/`14898108`/`14898109` completed upper and wide K4
   mean-by-type scale-grid follow-ups. The no-op upper grid `0.40/0.45/0.50`
   reaches 35.36%, the source-score upper grid reaches 35.30%, and the no-op
   wide grid `0.35/0.45/0.55` reaches 35.13%. Summary jobs `14898110`/
-  `14898111`/`14898112` and rebuild job `14898113` completed. The best clean row
-  remains the `0.35/0.40/0.45` no-op grid at 35.42%.
+  `14898111`/`14898112` and rebuild job `14898113` completed. At that stage the
+  best clean row remained the `0.35/0.40/0.45` no-op grid at 35.42%.
 - Job `14898293` completed the CPU Apptainer smoke for the residual action-L2
   penalty path with the best scale-grid/no-op configuration.
 - Jobs `14898327`/`14898329`/`14898331` completed minimum-energy tangent GPU