Clarify archived task curation rewards

seta_v2/README.md

@@ -2,7 +2,7 @@
 
 ## Dataset stats
 
-- **Tasks:** 4,567 native tasks; 1,414 omitted upstream, 46 excluded in v1. **3,107 v1 tasks → 2,205 kept**, including 55 newly retained after pass@16 recovery; **902 temporarily excluded**.
+- **Tasks:** 4,567 native tasks; 1,414 omitted upstream, 46 excluded in v1. **3,107 v1 tasks → 2,205 kept**, including 55 newly retained after pass@16 recovery; **902 archived from the current v2 release**.
 - **Images:** reuse **10 of v1's 11 cached Apptainer environments**. No additional images or task repairs in v2; original unique image count was not recorded.
 
 ## Dataset characteristics
@@ -30,14 +30,14 @@ The published task list combines the frozen 15 October 2026, 20:57 UTC snapshot
 | Qwen3-30B-A3B-Thinking | 1,459 |
 | Either model | **2,205** |
 
-| Still temporarily excluded | Tasks |
+| Archived from this release | Tasks |
 | --- | ---: |
 | Incomplete results, no success recorded | 509 |
 | Complete 16 attempts per model, no success | 364 |
 | No scored results | 29 |
 | Total | **902** |
 
-The recovery newly retained 55 tasks: 50 first successes for Instruct and 5 for Thinking. Another 194 existing tasks gained a first observed success for one model; model totals count each task once. Missing verifier rewards remain unknown, not failures.
+The recovery newly retained 55 tasks: 50 first successes for Instruct and 5 for Thinking. Another 194 existing tasks gained a first observed success for one model; model totals count each task once. The 509 tasks with incomplete results and no recorded success are archived from this release; each recorded failed trajectory receives curation reward 0. For attempts where the verifier produced no reward, the raw `verifier_reward` remains null and the error message is preserved. The 29 tasks with no scored results remain unscored.
 
 ## Successful attempts in the original snapshot
 

seta_v2/metadata/selection.json

@@ -202118,6 +202118,11 @@
       "no_success_in_complete_pass16": 364,
       "no_results_yet": 29
     },
-    "attempt_counts_note": "Original snapshot rewards and counts remain under each task model; recovery evidence is recorded separately. Do not sum attempt counts without deduplicating attempt IDs."
+    "attempt_counts_note": "Original snapshot rewards and counts remain under each task model; recovery evidence is recorded separately. Do not sum attempt counts without deduplicating attempt IDs.",
+    "archival_policy": {
+      "incomplete_results_no_success": "Archive the task from this release. Recorded failed trajectories receive curation reward 0.",
+      "missing_verifier_output": "Preserve raw verifier_reward as null with its error message; do not overwrite the raw result.",
+      "no_scored_results": "Keep archived and unscored until evidence is available."
+    }
   }
 }

seta_v2/runs/2026-10-15-pass16-recovery/run.json

@@ -925,5 +925,6 @@
       "qwen3-30b-thinking": 1459
     },
     "baseline_attempt_counts_preserved": true
-  }
+  },
+  "curation_policy": "For incomplete task results with recorded failed trajectories, archive the task from the current release and treat each recorded failed trajectory as curation reward 0. Raw verifier_reward remains null when no verifier output exists; preserve the original error message."
 }