Auto-sync: 2026-06-30 20:26:14

results/paper_analysis.json

@@ -1,6 +1,6 @@
 {
   "best_clean_key": "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmchallenger001",
-  "generated_utc": "2026-06-30T23:47:13+00:00",
+  "generated_utc": "2026-07-01T00:31:55+00:00",
   "mechanism_gap": {
     "best_clean_vs_direct_same_ckpt": 0.0776811594202898,
     "best_clean_vs_h16": 0.06318840579710144,
@@ -2305,6 +2305,16 @@
       "source": "results/h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_challenger0p01_norm_summary.json",
       "std_success": 0.015780472566743406
     },
+    "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin003_challenger001": {
+      "label": "K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper Pick/Pull margin 0.03",
+      "missing": true,
+      "source": "results/h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p03_challenger0p01_summary.json"
+    },
+    "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin005_challenger001": {
+      "label": "K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper Pick/Pull margin 0.05",
+      "missing": true,
+      "source": "results/h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p05_challenger0p01_summary.json"
+    },
     "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001": {
       "ci95_success": 0.028111505659259583,
       "label": "K4 compatible tangents, no-op bonus 0.03, near-miss/wrong-gripper challenger gate 0.01",
@@ -2499,6 +2509,200 @@
       "source": "results/h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmwgchallenger0p01_pickpull_norm_summary.json",
       "std_success": 0.01659024698116427
     },
+    "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin003_challenger001": {
+      "ci95_success": 0.03465460086179854,
+      "label": "K4 compatible tangents, no-op bonus 0.03, near-miss challenger 0.01, wrong-gripper margin 0.03",
+      "mean_action_mse_to_best": 0.42052025424397516,
+      "mean_progress": 0.5725739098747895,
+      "mean_success": 0.35768115942028983,
+      "num_completed": 3,
+      "per_task_success": {
+        "LiftPegUpright-v1": {
+          "mean_num_groups": 102.0,
+          "mean_success": 0.2564642064288538,
+          "std_success": 0.04855619832211069
+        },
+        "PickCube-v1": {
+          "mean_num_groups": 196.66666666666666,
+          "mean_success": 0.3289542583020844,
+          "std_success": 0.04030135158178743
+        },
+        "PullCube-v1": {
+          "mean_num_groups": 81.0,
+          "mean_success": 0.228271436166173,
+          "std_success": 0.03525140677829959
+        },
+        "PushCube-v1": {
+          "mean_num_groups": 102.0,
+          "mean_success": 0.757403601199683,
+          "std_success": 0.07781968814460603
+        },
+        "StackCube-v1": {
+          "mean_num_groups": 93.33333333333333,
+          "mean_success": 0.19336959336959336,
+          "std_success": 0.08682910965604235
+        }
+      },
+      "seed_action_mse_to_best": {
+        "0": 0.39712414266136675,
+        "1": 0.4136190785184179,
+        "2": 0.45081754155214065
+      },
+      "seed_progress": {
+        "0": 0.5605675064440331,
+        "1": 0.5702972661898188,
+        "2": 0.5868569569905167
+      },
+      "seed_success": {
+        "0": 0.3565217391304348,
+        "1": 0.3443478260869565,
+        "2": 0.37217391304347824
+      },
+      "selected_candidate_type_counts": {
+        "retrieval_residual_policy_residual": 1400,
+        "retrieval_residual_residual_near_miss": 91,
+        "retrieval_residual_residual_near_miss+residual_wrong_gripper": 5,
+        "retrieval_residual_residual_no_op": 38,
+        "retrieval_residual_residual_no_op+residual_wrong_gripper": 27,
+        "retrieval_residual_residual_wrong_gripper": 164
+      },
+      "selected_residual_scale_counts": {
+        "0.35": 1480,
+        "0.4": 44,
+        "0.45": 201
+      },
+      "selected_type_outcomes": {
+        "retrieval_residual_policy_residual": {
+          "count": 1400.0,
+          "mean_progress": 0.545508016923642,
+          "success_rate": 0.3171428571428571
+        },
+        "retrieval_residual_residual_near_miss": {
+          "count": 91.0,
+          "mean_progress": 0.5661385127926594,
+          "success_rate": 0.38461538461538464
+        },
+        "retrieval_residual_residual_near_miss+residual_wrong_gripper": {
+          "count": 5.0,
+          "mean_progress": 0.7557853102684021,
+          "success_rate": 0.6
+        },
+        "retrieval_residual_residual_no_op": {
+          "count": 38.0,
+          "mean_progress": 0.7757954887653652,
+          "success_rate": 0.5263157894736842
+        },
+        "retrieval_residual_residual_no_op+residual_wrong_gripper": {
+          "count": 27.0,
+          "mean_progress": 0.6916872307106301,
+          "success_rate": 0.48148148148148145
+        },
+        "retrieval_residual_residual_wrong_gripper": {
+          "count": 164.0,
+          "mean_progress": 0.7349113159949278,
+          "success_rate": 0.6219512195121951
+        }
+      },
+      "source": "results/h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmwgmargin0p03_challenger0p01_summary.json",
+      "std_success": 0.01394922830784459
+    },
+    "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin005_challenger001": {
+      "ci95_success": 0.03242840579710147,
+      "label": "K4 compatible tangents, no-op bonus 0.03, near-miss challenger 0.01, wrong-gripper margin 0.05",
+      "mean_action_mse_to_best": 0.41712138921886255,
+      "mean_progress": 0.5713764009092921,
+      "mean_success": 0.35594202898550725,
+      "num_completed": 3,
+      "per_task_success": {
+        "LiftPegUpright-v1": {
+          "mean_num_groups": 102.0,
+          "mean_success": 0.2564642064288538,
+          "std_success": 0.04855619832211069
+        },
+        "PickCube-v1": {
+          "mean_num_groups": 196.66666666666666,
+          "mean_success": 0.3272707566185827,
+          "std_success": 0.03916710518683381
+        },
+        "PullCube-v1": {
+          "mean_num_groups": 81.0,
+          "mean_success": 0.22388547125389233,
+          "std_success": 0.04131573991107309
+        },
+        "PushCube-v1": {
+          "mean_num_groups": 102.0,
+          "mean_success": 0.7541032711666796,
+          "std_success": 0.07644139692615382
+        },
+        "StackCube-v1": {
+          "mean_num_groups": 93.33333333333333,
+          "mean_success": 0.19336959336959336,
+          "std_success": 0.08682910965604235
+        }
+      },
+      "seed_action_mse_to_best": {
+        "0": 0.3951253530658458,
+        "1": 0.411393547800734,
+        "2": 0.44484526679000774
+      },
+      "seed_progress": {
+        "0": 0.5606453491382949,
+        "1": 0.5687696125099192,
+        "2": 0.5847142410796622
+      },
+      "seed_success": {
+        "0": 0.3565217391304348,
+        "1": 0.3426086956521739,
+        "2": 0.36869565217391304
+      },
+      "selected_candidate_type_counts": {
+        "retrieval_residual_policy_residual": 1447,
+        "retrieval_residual_residual_near_miss": 91,
+        "retrieval_residual_residual_near_miss+residual_wrong_gripper": 5,
+        "retrieval_residual_residual_no_op": 38,
+        "retrieval_residual_residual_no_op+residual_wrong_gripper": 27,
+        "retrieval_residual_residual_wrong_gripper": 117
+      },
+      "selected_residual_scale_counts": {
+        "0.35": 1507,
+        "0.4": 39,
+        "0.45": 179
+      },
+      "selected_type_outcomes": {
+        "retrieval_residual_policy_residual": {
+          "count": 1447.0,
+          "mean_progress": 0.5467335695066794,
+          "success_rate": 0.3213545266067726
+        },
+        "retrieval_residual_residual_near_miss": {
+          "count": 91.0,
+          "mean_progress": 0.5661385127926594,
+          "success_rate": 0.38461538461538464
+        },
+        "retrieval_residual_residual_near_miss+residual_wrong_gripper": {
+          "count": 5.0,
+          "mean_progress": 0.7557853102684021,
+          "success_rate": 0.6
+        },
+        "retrieval_residual_residual_no_op": {
+          "count": 38.0,
+          "mean_progress": 0.7757954887653652,
+          "success_rate": 0.5263157894736842
+        },
+        "retrieval_residual_residual_no_op+residual_wrong_gripper": {
+          "count": 27.0,
+          "mean_progress": 0.6916872307106301,
+          "success_rate": 0.48148148148148145
+        },
+        "retrieval_residual_residual_wrong_gripper": {
+          "count": 117.0,
+          "mean_progress": 0.7781837732873411,
+          "success_rate": 0.6666666666666666
+        }
+      },
+      "source": "results/h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmwgmargin0p05_challenger0p01_summary.json",
+      "std_success": 0.01305313652080894
+    },
     "residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_oraclek8": {
       "candidate_oracle_rollouts": 8,
       "candidate_oracle_type_counts": {

results/paper_analysis.md

@@ -1,6 +1,6 @@
 # Paper Analysis
 
-Generated: `2026-06-30T23:47:13+00:00`
+Generated: `2026-07-01T00:31:55+00:00`
 
 ## Main Seed Statistics
 
@@ -61,8 +61,12 @@ Generated: `2026-06-30T23:47:13+00:00`
 | residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmchallenger001_scales035040 | K4 compatible tangents, no-op bonus 0.03, near-miss challenger gate 0.01, scale-gated 0.35/0.40 | 3 | 36.00% +/- 1.31 | +/- 3.26 | 57.37% | 0.407 | +6.26 pp |
 | residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmchallenger003 | K4 compatible tangents, no-op bonus 0.03, near-miss challenger gate 0.03 | 3 | 35.94% +/- 1.28 | +/- 3.18 | 57.36% | 0.407 | +6.20 pp |
 | residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001 | K4 compatible tangents, no-op bonus 0.03, near-miss/wrong-gripper challenger gate 0.01 | 3 | 35.94% +/- 1.13 | +/- 2.81 | 57.40% | 0.424 | +6.20 pp |
+| residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin003_challenger001 | K4 compatible tangents, no-op bonus 0.03, near-miss challenger 0.01, wrong-gripper margin 0.03 | 3 | 35.77% +/- 1.39 | +/- 3.47 | 57.26% | 0.421 | +6.03 pp |
+| residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin005_challenger001 | K4 compatible tangents, no-op bonus 0.03, near-miss challenger 0.01, wrong-gripper margin 0.05 | 3 | 35.59% +/- 1.31 | +/- 3.24 | 57.14% | 0.417 | +5.86 pp |
 | residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001_pickpull | K4 compatible tangents, no-op bonus 0.03, near-miss/wrong-gripper challenger gate 0.01 on Pick/Pull only | 3 | 36.00% +/- 1.66 | +/- 4.12 | 57.36% | 0.418 | +6.26 pp |
 | residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_challenger001 | K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper challenger on Pick/Pull, gate 0.01 | 3 | 36.06% +/- 1.58 | +/- 3.92 | 57.41% | 0.419 | +6.32 pp |
+| residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin003_challenger001 | K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper Pick/Pull margin 0.03 | 0 | missing | missing | missing | missing | missing |
+| residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin005_challenger001 | K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper Pick/Pull margin 0.05 | 0 | missing | missing | missing | missing | missing |
 | typed_proposal_lattice_types6_prepend_margin000 | Typed proposal lattice head, six families, policy-prepended margin 0.00 | 3 | 31.30% +/- 3.81 | +/- 9.48 | 54.03% | 0.480 | +1.57 pp |
 | typed_proposal_lattice_types6_prepend_margin005 | Typed proposal lattice head, six families, policy-prepended margin 0.05 | 3 | 32.17% +/- 3.05 | +/- 7.57 | 54.74% | 0.467 | +2.43 pp |
 | typed_proposal_lattice_types4safe_prepend_margin005 | Typed proposal lattice head, safe four families, policy-prepended margin 0.05 | 3 | 33.45% +/- 3.39 | +/- 8.43 | 55.47% | 0.456 | +3.71 pp |

results/paper_table_status.json

@@ -1562,6 +1562,44 @@
       "best_config": null,
       "gain_vs_h16_policy": 0.06202898550724639
     },
+    {
+      "key": "retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin003_challenger001",
+      "label": "K4 compatible residual retrieval, near-miss challenger 0.01, wrong-gripper margin 0.03",
+      "path": "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmwgmargin0p03_challenger0p01_summary.json",
+      "clean_deployment": "yes",
+      "same_state_proposals": "no",
+      "expert_proposal": "no",
+      "story_role": "family-specific reliability margin: keep near-miss sensitive while requiring stronger wrong-gripper evidence",
+      "fallback_success": null,
+      "pending_job": "14989648/14989649",
+      "path_exists": true,
+      "status": "complete",
+      "success": 0.35768115942028983,
+      "std_success": 0.013949228307844588,
+      "completed_seeds": null,
+      "num_completed": 3,
+      "best_config": null,
+      "gain_vs_h16_policy": 0.06028985507246376
+    },
+    {
+      "key": "retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin005_challenger001",
+      "label": "K4 compatible residual retrieval, near-miss challenger 0.01, wrong-gripper margin 0.05",
+      "path": "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmwgmargin0p05_challenger0p01_summary.json",
+      "clean_deployment": "yes",
+      "same_state_proposals": "no",
+      "expert_proposal": "no",
+      "story_role": "family-specific reliability margin upper setting for wrong-gripper challenger evidence",
+      "fallback_success": null,
+      "pending_job": "14989651/14989659",
+      "path_exists": true,
+      "status": "complete",
+      "success": 0.35594202898550725,
+      "std_success": 0.01305313652080894,
+      "completed_seeds": null,
+      "num_completed": 3,
+      "best_config": null,
+      "gain_vs_h16_policy": 0.058550724637681184
+    },
     {
       "key": "retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001_pickpull",
       "label": "K4 compatible residual retrieval, near-miss/wrong-gripper challenger gate 0.01 on Pick/Pull only",
@@ -1600,6 +1638,44 @@
       "best_config": null,
       "gain_vs_h16_policy": 0.0631884057971015
     },
+    {
+      "key": "retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin003_challenger001",
+      "label": "K4 compatible residual retrieval, near-miss global + wrong-gripper Pick/Pull margin 0.03",
+      "path": "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p03_challenger0p01_summary.json",
+      "clean_deployment": "yes",
+      "same_state_proposals": "no",
+      "expert_proposal": "no",
+      "story_role": "type-specific affordance margin: keep near-miss globally sensitive while requiring stronger wrong-gripper evidence only on Pick/Pull",
+      "fallback_success": null,
+      "pending_job": "14989757/14989775",
+      "path_exists": false,
+      "status": "pending",
+      "success": null,
+      "std_success": null,
+      "completed_seeds": null,
+      "num_completed": null,
+      "best_config": null,
+      "gain_vs_h16_policy": null
+    },
+    {
+      "key": "retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin005_challenger001",
+      "label": "K4 compatible residual retrieval, near-miss global + wrong-gripper Pick/Pull margin 0.05",
+      "path": "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p05_challenger0p01_summary.json",
+      "clean_deployment": "yes",
+      "same_state_proposals": "no",
+      "expert_proposal": "no",
+      "story_role": "upper setting for type-specific wrong-gripper evidence on Pick/Pull while preserving global near-miss calibration",
+      "fallback_success": null,
+      "pending_job": "14989760/14989780",
+      "path_exists": false,
+      "status": "pending",
+      "success": null,
+      "std_success": null,
+      "completed_seeds": null,
+      "num_completed": null,
+      "best_config": null,
+      "gain_vs_h16_policy": null
+    },
     {
       "key": "typed_proposal_lattice_types6_prepend_margin000",
       "label": "Typed proposal lattice head, six families, policy-prepended margin 0.00",

results/paper_table_status.md

@@ -85,8 +85,12 @@ Baseline h=16 policy: 29.74%
 | retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmchallenger001_scales035040 | K4 compatible residual retrieval, near-miss challenger gate 0.01, scale-gated 0.35/0.40 | complete | 36.00% | +6.26 pp | yes | no | no | trace-motivated near-miss challenger with tangent-length reliability gating over the two shortest scales |
 | retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmchallenger003 | K4 compatible residual retrieval, near-miss challenger gate 0.03 | complete | 35.94% | +6.20 pp | yes | no | no | upper-margin sensitivity for trace-motivated singleton near-miss challenger calibration |
 | retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001 | K4 compatible residual retrieval, near-miss/wrong-gripper challenger gate 0.01 | complete | 35.94% | +6.20 pp | yes | no | no | trace-motivated challenger calibration test for whether wrong-gripper residuals carry conditional selector headroom |
+| retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin003_challenger001 | K4 compatible residual retrieval, near-miss challenger 0.01, wrong-gripper margin 0.03 | complete | 35.77% | +6.03 pp | yes | no | no | family-specific reliability margin: keep near-miss sensitive while requiring stronger wrong-gripper evidence |
+| retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin005_challenger001 | K4 compatible residual retrieval, near-miss challenger 0.01, wrong-gripper margin 0.05 | complete | 35.59% | +5.86 pp | yes | no | no | family-specific reliability margin upper setting for wrong-gripper challenger evidence |
 | retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001_pickpull | K4 compatible residual retrieval, near-miss/wrong-gripper challenger gate 0.01 on Pick/Pull only | complete | 36.00% | +6.26 pp | yes | no | no | task-conditional challenger gate using branch trace evidence that wrong-gripper tangents help Pick/Pull but harm other tasks |
 | retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_challenger001 | K4 compatible residual retrieval, near-miss global + wrong-gripper challenger on Pick/Pull, gate 0.01 | complete | 36.06% | +6.32 pp | yes | no | no | type-specific affordance gate: keep near-miss global while restricting wrong-gripper challenger to acquisition/contact tasks |
+| retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin003_challenger001 | K4 compatible residual retrieval, near-miss global + wrong-gripper Pick/Pull margin 0.03 | pending 14989757/14989775 | pending | pending | yes | no | no | type-specific affordance margin: keep near-miss globally sensitive while requiring stronger wrong-gripper evidence only on Pick/Pull |
+| retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin005_challenger001 | K4 compatible residual retrieval, near-miss global + wrong-gripper Pick/Pull margin 0.05 | pending 14989760/14989780 | pending | pending | yes | no | no | upper setting for type-specific wrong-gripper evidence on Pick/Pull while preserving global near-miss calibration |
 | typed_proposal_lattice_types6_prepend_margin000 | Typed proposal lattice head, six families, policy-prepended margin 0.00 | complete | 31.30% | +1.57 pp | yes | no | no | model-generated counterfactual proposal support test; field scores typed proposals rather than retrieved train-state residuals |
 | typed_proposal_lattice_types6_prepend_margin005 | Typed proposal lattice head, six families, policy-prepended margin 0.05 | complete | 32.17% | +2.43 pp | yes | no | no | model-generated counterfactual proposal support test with a conservative policy-abstention margin |
 | typed_proposal_lattice_types4safe_prepend_margin005 | Typed proposal lattice head, safe four families, policy-prepended margin 0.05 | complete | 33.45% | +3.71 pp | yes | no | no | anti-goal proposal-family mask test for the model-generated typed lattice |

scripts/build_paper_analysis.py

@@ -497,6 +497,22 @@ METHODS = [
             "k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_challenger0p01_norm_summary.json"
         ),
     ),
+    MethodSpec(
+        key="residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin003_challenger001",
+        label="K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper Pick/Pull margin 0.03",
+        summary_path=(
+            "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_"
+            "k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p03_challenger0p01_summary.json"
+        ),
+    ),
+    MethodSpec(
+        key="residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin005_challenger001",
+        label="K4 compatible tangents, no-op bonus 0.03, near-miss global + wrong-gripper Pick/Pull margin 0.05",
+        summary_path=(
+            "h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_"
+            "k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p05_challenger0p01_summary.json"
+        ),
+    ),
     MethodSpec(
         key="typed_proposal_lattice_types6_prepend_margin000",
         label="Typed proposal lattice head, six families, policy-prepended margin 0.00",

scripts/build_paper_table_status.py

@@ -843,7 +843,7 @@ SPECS = [
         same_state_proposals="no",
         expert_proposal="no",
         story_role="family-specific reliability margin: keep near-miss sensitive while requiring stronger wrong-gripper evidence",
-        pending_job="pending",
+        pending_job="14989648/14989649",
     ),
     ResultSpec(
         key="retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgmargin005_challenger001",
@@ -853,7 +853,7 @@ SPECS = [
         same_state_proposals="no",
         expert_proposal="no",
         story_role="family-specific reliability margin upper setting for wrong-gripper challenger evidence",
-        pending_job="pending",
+        pending_job="14989651/14989659",
     ),
     ResultSpec(
         key="retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmwgchallenger001_pickpull",
@@ -875,6 +875,26 @@ SPECS = [
         story_role="type-specific affordance gate: keep near-miss global while restricting wrong-gripper challenger to acquisition/contact tasks",
         pending_job="14974610/14974611",
     ),
+    ResultSpec(
+        key="retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin003_challenger001",
+        label="K4 compatible residual retrieval, near-miss global + wrong-gripper Pick/Pull margin 0.03",
+        path="h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p03_challenger0p01_summary.json",
+        clean_deployment="yes",
+        same_state_proposals="no",
+        expert_proposal="no",
+        story_role="type-specific affordance margin: keep near-miss globally sensitive while requiring stronger wrong-gripper evidence only on Pick/Pull",
+        pending_job="14989757/14989775",
+    ),
+    ResultSpec(
+        key="retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_noopbonus003_nmglobal_wgpickpull_wgmargin005_challenger001",
+        label="K4 compatible residual retrieval, near-miss global + wrong-gripper Pick/Pull margin 0.05",
+        path="h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_k4_composemasked_dropnmnoop_grid035040045_safe_margin0p20_noopbonus0p03_nmglobal_wgpickpull_wgmargin0p05_challenger0p01_summary.json",
+        clean_deployment="yes",
+        same_state_proposals="no",
+        expert_proposal="no",
+        story_role="upper setting for type-specific wrong-gripper evidence on Pick/Pull while preserving global near-miss calibration",
+        pending_job="14989760/14989780",
+    ),
     ResultSpec(
         key="typed_proposal_lattice_types6_prepend_margin000",
         label="Typed proposal lattice head, six families, policy-prepended margin 0.00",