Publish partial CalibForge v2 and verified SETA and CalibForge traces

calibforge_v2/README.md

@@ -0,0 +1,75 @@
+# CalibForge v2
+
+## Dataset stats
+
+- **Tasks:** 5,431 original; 5,317 retained in v1. **84 retained in v2**, all previously patched; **5,233 temporarily excluded**. The 114 earlier exclusions remain in v1.
+- **Images:** reuse **84 cached v1 Apptainer environments**; no new image builds or task repairs in v2. V1 has 5,210 cached environments.
+
+## Dataset characteristics
+
+<details><summary>Domains</summary>
+
+Terminal tasks spanning debugging, software engineering, data processing, data science, security, mathematics and optimization.
+
+</details>
+<details><summary>Capabilities</summary>
+
+Exploration, specification adherence and error recovery through inspecting files, modifying programs and checking executable outputs.
+
+</details>
+
+Intended benchmark transfer: Terminal-Bench 2.0 and OpenThoughts-TBLite 2.0; data-science tasks also overlap DS-1000.
+
+## Partial-result selection
+
+Snapshot: **15 October 2026, 21:02 UTC**. Qwen3-30B-A3B-Instruct-2507, with eight planned attempts per task.
+
+**Of 5,317 tasks, 209 are complete, 97 have started but are incomplete, and 5,011 have no recorded attempt.** Complete means all eight attempts have a recorded score.
+
+**84 tasks are solved at least once:** 66 with complete results and 18 with partial results. A single successful attempt is sufficient for retention.
+
+| Temporary exclusion reason | Tasks |
+| --- | ---: |
+| Complete, no success | 143 |
+| Started but incomplete, no success yet | 79 |
+| No recorded attempt | 5,011 |
+| Total | **5,233** |
+
+Missing results remain unknown, not failures. These exclusions can be reconsidered as runs finish.
+
+Build, no-op and rule-based checks carry over from [v1](../Calibforge_environment-reliable/README.md); no reference solutions are provided. V2 preserves v1 payloads. For all 84 retained tasks, the evaluated variant differs only by the added timeout/anti-cheating instruction sentence and trailing instruction whitespace; all other file bytes match. The variant has not been rerun with that sentence.
+
+## Successful attempts among retained tasks
+
+Observed counts from up to eight attempts, not complete pass@8 estimates.
+
+| Instruct successes | Tasks |
+| --- | ---: |
+| 1 | 33 |
+| 2 | 17 |
+| 3 | 10 |
+| 4 | 8 |
+| 5 | 4 |
+| 6 | 2 |
+| 7 | 4 |
+| 8 | 6 |
+
+## Files and selecting tasks
+
+`tasks.parquet` contains retained payloads, `instruct_successes`, `instruct_graded_attempts`, `instruct_missing_attempts` and `target_attempts_per_model` (8). `assets/selection.json` records all task statuses, attempt IDs, rewards and the pinned v1 source. Excluded payloads remain in v1; no duplicate archive Parquet is uploaded. `patch.py` reproduces this snapshot.
+
+```python
+import pyarrow.parquet as pq
+subset = pq.read_table("tasks.parquet", filters=[("instruct_successes", "=", 1)])
+pq.write_table(subset, "selected.parquet", compression="zstd")
+```
+
+## Traces
+
+[Compressed trace Parquet](traces/instruct.parquet) contains **2,013 recorded attempts**, successful and failed, from retained and excluded tasks. All conversations are present. Six attempts have no recorded score; their rewards remain null. These traces reproduce the snapshot's 84 solved tasks.
+
+Each row includes task ID, evaluated content hash, model, attempt ID, snapshot reward, raw verifier reward, retention flag and full `trajectory` and `result` JSON. No traces are invented for unfinished or unstarted attempts.
+
+## Runtime
+
+Apptainer **1.5.4**, Python **3.12.14**, Harbor **0.8.1**; [Marin Harbor bridge](https://github.com/marin-community/harbor/tree/7faf878c14b6d72737579ec936ebf5f74ba3194d) and [runtime patches](https://github.com/MLI-lab/OT-next-data/tree/main/harbor_patches).

calibforge_v2/patch.py

@@ -0,0 +1,21 @@
+"""Reproduce CalibForge v2 from pinned v1 and a frozen partial-results snapshot."""
+import argparse,json,hashlib
+from pathlib import Path
+import pyarrow as pa,pyarrow.parquet as pq
+from huggingface_hub import hf_hub_download
+p=argparse.ArgumentParser(description=__doc__);p.add_argument('--out',type=Path,required=True);p.add_argument('--source',type=Path);a=p.parse_args();a.out.mkdir(parents=True,exist_ok=True)
+assert not (a.out/'tasks.parquet').exists()
+m=json.loads((Path(__file__).parent/'assets/selection.json').read_text());f=a.source or Path(hf_hub_download(m['repo_id'],m['source_folder']+'/tasks.parquet',revision=m['source_revision'],repo_type='dataset',cache_dir=a.out/'.source-cache'));assert hashlib.sha256(f.read_bytes()).hexdigest()==m['source_sha256']
+src=pq.ParquetFile(f);schema=src.schema_arrow.remove_metadata()
+for name in ['instruct_successes','instruct_graded_attempts','instruct_missing_attempts','target_attempts_per_model']:schema=schema.append(pa.field(name,pa.int64()))
+rows=[]
+for batch in src.iter_batches(batch_size=32):
+ for r in batch.to_pylist():
+  s=m['tasks'][r['path']];assert r['content_sha256']==s['content_sha256']
+  if not s['kept']:continue
+  for k in ['instruct_successes','instruct_graded_attempts','instruct_missing_attempts']:r[k]=s[k]
+  assert r['instruct_successes']>0;r['target_attempts_per_model']=8;rows.append(r)
+assert len(rows)==84
+names={r['path'] for r in rows};im=json.loads(src.schema_arrow.metadata[b'ot.images.v1']);im['tasks']={k:v for k,v in im['tasks'].items() if k in names};assert set(im['tasks'])==names
+bundles={e['bundle'] for v in im['tasks'].values() for e in v['environments']};im['bundles']={k:v for k,v in im['bundles'].items() if k in bundles};im['missing']=[]
+meta={b'ot.images.v1':json.dumps(im).encode(),b'ot.release.v2':json.dumps({k:v for k,v in m.items() if k!='tasks'}).encode()};pq.write_table(pa.Table.from_pylist(rows,schema=schema).replace_schema_metadata(meta),a.out/'tasks.parquet',compression='zstd');print('Retained',len(rows),'images reused',len(bundles))

calibforge_v2/tasks.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:736549e9f301bd3ecd9c1760624b090f67e7077ca3de635c1a4d3ab305d5f86b
+size 1956215

calibforge_v2/traces/instruct.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4698090133fe24d248ed6c2b5be2a95490ea77e5e40d76db98627b3064ac44b7
+size 20751029

calibforge_v2/traces/manifest.json

@@ -0,0 +1,9 @@
+{
+  "snapshot_utc": "2026-10-15T21:02:40.200662+00:00",
+  "attempts": 2013,
+  "missing_traces": 0,
+  "ungraded_attempts": 6,
+  "successful_attempts": 237,
+  "solved_tasks": 84,
+  "sha256": "4698090133fe24d248ed6c2b5be2a95490ea77e5e40d76db98627b3064ac44b7"
+}

seta_v2/README.md

@@ -77,7 +77,11 @@ subset = pq.read_table("tasks.parquet", filters=[("instruct_successes", "=", 1)]
 pq.write_table(subset, "selected.parquet", compression="zstd")
 ```
 
-Traces are not included in this snapshot.
+## Traces
+
+[Compressed trace Parquets](traces) contain **91,898 recorded attempts**, including successful attempts, failed attempts and retries, from retained and excluded tasks. They reproduce the snapshot's per-task graded and successful attempt counts. **165 conversations are missing** and marked `trajectory_found=false`; **4,746 attempts have no snapshot score** and retain null rewards.
+
+Each row includes task ID, content hash, model, attempt ID, snapshot reward, raw verifier reward, retention flag and full `trajectory` and `result` JSON. Additional ungraded retry records are preserved for auditing and do not increase the published graded-attempt counts.
 
 ## Runtime
 

seta_v2/traces/000.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4670c1c72c5aa74aeac0fb8edde66e8d28b4562b98dbb864f3097d179c5ebf54
+size 2144724

seta_v2/traces/001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e964a65ddc7e670b455db69dfb6b7d4960950bc9d588bca7f88076a34aef68b
+size 1841391

seta_v2/traces/002.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d9e8ef5e7d33553a08b8aae0ee4e3f43b01214e1f7af3a814319f488680710a
+size 1295031

seta_v2/traces/003.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e41ba43e3731bec0cb7ef7f99762ca42923d06d9bb0ff8cf3b9ca80fd1198a3
+size 1287198

seta_v2/traces/004.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d7e488d33b5ac24efff764b73378072d86eb5880d0dc75dced8c4f1338dfe19
+size 1692865

seta_v2/traces/005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e59e5b1e3b3b8e7d489338c2c08fda4b10c04236ea0c16181f978de93d449e65
+size 1836328

seta_v2/traces/006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d0c94cca6998bff785752940b70d9fde41cf4f06f65b040d7b7933ace43fb98
+size 2573437

seta_v2/traces/007.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a47b2526ed9debcb5ced3251859b09382195f9e4655fd5863766978edae576af
+size 2599527

seta_v2/traces/008.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e58404e7178ce7fb44f6e7652b4a12a783bf266f609fa93dc7ff000fbb698628
+size 1758508

seta_v2/traces/009.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b6cd323c8b2fb62c9afe7d266c75436dec05090fc855037778173db869ef8aa
+size 1139776

seta_v2/traces/010.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb630bcb1053f5f7c74dc2a54afdc25b7093f9a14c51e272081fb1a1eeecde25
+size 1370683

seta_v2/traces/011.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dede2b34133ba719412915bc258fdc927f9c047df8163bde8cb1a82c94483fc2
+size 1511759

seta_v2/traces/012.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7687252e9a86b7e7194773146016573e97e8732589378fb23c528a7c3265e53
+size 1929351

seta_v2/traces/013.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a68fa7d96526ee1dc4f9f016b81f9368b6bf0c9afcf59a7f63846b12fca3845
+size 1382068

seta_v2/traces/014.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2d680f62c70cad3ef25578adb7d088518daed9fd82fbd8febd5a59f82cd36521
+size 2223328

seta_v2/traces/015.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:28f5c5760c8fcb5578d912ae8718ef1e0bcc70dfd174ae418b3d976e75e230d9
+size 968269

seta_v2/traces/016.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ee86ba4255809b17577adbd2499b6bc25c401183658b6154f9b51d5248c13859
+size 8095756

seta_v2/traces/017.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48049b68aedc459ff113dab70f5e6bb5f0616e21555d1b735d9d133e5275a716
+size 8519047

seta_v2/traces/018.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:18ee48bf88a458b0bb98e148c5ced73fb0a0459a25a25d2026ac2e2d18dd8940
+size 4904941

seta_v2/traces/019.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8ce8d61aeb4c0b5fc199d83a5d962e30f6de0cb605544b0c43d1c5c60b562e3
+size 4536895

seta_v2/traces/020.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e0eaa33ca696ccead8b158dc0c47ca1ebcd2ad5b7d302ad82ab0c70adec5061
+size 7519873

seta_v2/traces/021.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b961042524e7e71402b3c3ed1501e3836a6d8874e73281a20d2a38a91333f5d
+size 7997704

seta_v2/traces/022.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e67415f1ab5a2394cbe34aa0825f9b2e2f1b393098670201c8752dddd37be35
+size 10519949

seta_v2/traces/023.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1940337ac54dbe7f759932944a2bf83f94b0ad6cdb70e448bc38426277d275d2
+size 3430583

seta_v2/traces/024.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92321d626ca8f7221223f672301ea23413658dd5cc249cc19adc1b5a270bcdcd
+size 7169049

seta_v2/traces/025.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c543b41d10df68bfbe623ba7ae50e8dcf3c08005ab21c23104fb70a9b11f1f93
+size 1473567

seta_v2/traces/026.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5214bcd927c9ec47e743343b6a254e146142f31d4168c4374e249420449c284
+size 1414997

seta_v2/traces/027.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6fabf10ea49159cad350eb58df7d0d8447b32f41b4a55ac87c2b83667db7e25c
+size 1209418

seta_v2/traces/028.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6583450cdaac28c21e6a093057f1da6c56347f1b39f6ac255d883c7874481ca
+size 2534370

seta_v2/traces/029.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b725fca41e11202586b1bb61b2d1089c29c8865cc835e7b8131f0d6918ea6be3
+size 2473067

seta_v2/traces/030.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:00a66ecb46547aa35402aaedf635abb1dcb1edffde65dcbdb6c1b6c112cdc08d
+size 1429958

seta_v2/traces/031.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1f966ce768fb1600d56162ff59e6bee37791bfa662d9b9e3d6b7686a5c816060
+size 1788344

seta_v2/traces/032.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a19556db57f089e8a376dcb70178412fa2b9a1f662f03ef82a36846a42a27f9a
+size 1253022

seta_v2/traces/033.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16c9a76f292ec76c19be4c912a747a2a128214c6a52f8a5bdcc4a349d9371ab7
+size 1384059

seta_v2/traces/034.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99f764235afd7f2be5a7d596c4a7fd7854f29a78d91a1b6ed326ab1e499a95d4
+size 1700559

seta_v2/traces/035.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:819b2d1b8dfee7f79601234bb6b2540fffa2923b5f714a40585dcce497b654b3
+size 1923459

seta_v2/traces/036.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:57306ef56e1602b44b22e04b39fed159b3766ecc9a63abd7fa8cabb7c05331ad
+size 2213625

seta_v2/traces/037.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d02f2e089b01fb175e354d36890c8f47ee80ca8e514a8baf14b13a696e29d901
+size 2531418

seta_v2/traces/038.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ae28ecf0147047b18316c4bbacb9645b3acf9cd33d52676ba093f6739031c37
+size 1844829

seta_v2/traces/039.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ac26d999aebfb1b282865f9f43e238848a4564d6700f7a020616c91bff7684e
+size 1947207

seta_v2/traces/040.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f984f90bcdf056b56b1aa45eb1ef516ef8ae6b05e3feadee2c640e75a3a2e39d
+size 1577599

seta_v2/traces/041.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19a84711954ba3dbc5fab363427aa002e39b89ca8e069e653cabe208c982c02a
+size 1708846

seta_v2/traces/042.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3eefd70c799dffee26f5971831e947d3343cf784dab7d153515ce5c3df7f3b83
+size 1281570