JLiangHe commited on
Commit
b2e6160
·
verified ·
1 Parent(s): d42dc63

Publish finished Granite, Bolt, Cascade, archived Muse and SFT SciCoding rerun

Browse files

PVC-direct full trajectory upload. HLE subsets combined; null judgments and explicit adjudications preserved. Active SFT TB and queued Granite repeat excluded.

This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +40 -0
  2. AGENTIC_CONFIG_ALIGNMENT.md +4 -0
  3. NVIDIA-Nemotron-3.5-Bolt-E4B/README.md +18 -0
  4. NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/events.jsonl.gz +3 -0
  5. NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/judge_trajectories.jsonl +0 -0
  6. NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/summary.json +1 -0
  7. NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/trajectories.jsonl +3 -0
  8. NVIDIA-Nemotron-3.5-Bolt-E4B/configs.json +0 -0
  9. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/events.jsonl.gz +3 -0
  10. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/judge_trajectories.jsonl +3 -0
  11. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/summary.json +1 -0
  12. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/trajectories.jsonl +3 -0
  13. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/events.jsonl.gz +3 -0
  14. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/judge_trajectories.jsonl +3 -0
  15. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/summary.json +1 -0
  16. NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/trajectories.jsonl +3 -0
  17. NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/events.jsonl.gz +3 -0
  18. NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/judge_trajectories.jsonl +3 -0
  19. NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/summary.json +1 -0
  20. NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/trajectories.jsonl +3 -0
  21. NVIDIA-Nemotron-3.5-Bolt-E4B/manifest.json +1 -0
  22. NVIDIA-Nemotron-3.5-Bolt-E4B/provenance.json +1 -0
  23. NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/judge_trajectories.jsonl +0 -0
  24. NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/native_artifacts.tar.gz +3 -0
  25. NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/summary.json +1 -0
  26. NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/trajectories.jsonl +3 -0
  27. NVIDIA-Nemotron-3.5-Bolt-E4B/summary.json +1 -0
  28. NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/judge_trajectories.jsonl +3 -0
  29. NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/native_artifacts.tar.gz +3 -0
  30. NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/summary.json +1 -0
  31. NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/trajectories.jsonl +3 -0
  32. Nemotron-Cascade-2-30B-A3B/README.md +18 -0
  33. Nemotron-Cascade-2-30B-A3B/arxivmath/events.jsonl.gz +3 -0
  34. Nemotron-Cascade-2-30B-A3B/arxivmath/judge_trajectories.jsonl +0 -0
  35. Nemotron-Cascade-2-30B-A3B/arxivmath/summary.json +1 -0
  36. Nemotron-Cascade-2-30B-A3B/arxivmath/trajectories.jsonl +3 -0
  37. Nemotron-Cascade-2-30B-A3B/configs.json +0 -0
  38. Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/events.jsonl.gz +3 -0
  39. Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/judge_trajectories.jsonl +3 -0
  40. Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/summary.json +1 -0
  41. Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/trajectories.jsonl +3 -0
  42. Nemotron-Cascade-2-30B-A3B/frontierscience-research/events.jsonl.gz +3 -0
  43. Nemotron-Cascade-2-30B-A3B/frontierscience-research/judge_trajectories.jsonl +3 -0
  44. Nemotron-Cascade-2-30B-A3B/frontierscience-research/summary.json +1 -0
  45. Nemotron-Cascade-2-30B-A3B/frontierscience-research/trajectories.jsonl +3 -0
  46. Nemotron-Cascade-2-30B-A3B/hle-verified/events.jsonl.gz +3 -0
  47. Nemotron-Cascade-2-30B-A3B/hle-verified/judge_trajectories.jsonl +3 -0
  48. Nemotron-Cascade-2-30B-A3B/hle-verified/summary.json +1 -0
  49. Nemotron-Cascade-2-30B-A3B/hle-verified/trajectories.jsonl +3 -0
  50. Nemotron-Cascade-2-30B-A3B/manifest.json +1 -0
.gitattributes CHANGED
@@ -240,3 +240,43 @@ backup/gpt-oss-120b/frontierscience-olympiad/judge_trajectories.jsonl filter=lfs
240
  backup/gpt-oss-120b/frontierscience-olympiad/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
241
  backup/gpt-oss-120b/frontierscience-research/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
242
  backup/gpt-oss-120b/frontierscience-research/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
240
  backup/gpt-oss-120b/frontierscience-olympiad/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
241
  backup/gpt-oss-120b/frontierscience-research/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
242
  backup/gpt-oss-120b/frontierscience-research/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
243
+ NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
244
+ NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
245
+ NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
246
+ NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
247
+ NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
248
+ NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
249
+ NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
250
+ NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
251
+ NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
252
+ NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
253
+ Nemotron-Cascade-2-30B-A3B/arxivmath/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
254
+ Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
255
+ Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
256
+ Nemotron-Cascade-2-30B-A3B/frontierscience-research/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
257
+ Nemotron-Cascade-2-30B-A3B/frontierscience-research/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
258
+ Nemotron-Cascade-2-30B-A3B/hle-verified/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
259
+ Nemotron-Cascade-2-30B-A3B/hle-verified/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
260
+ Nemotron-Cascade-2-30B-A3B/scicoding-agent/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
261
+ Nemotron-Cascade-2-30B-A3B/terminal-bench-2.1/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
262
+ Nemotron-Cascade-2-30B-A3B/terminal-bench-2.1/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
263
+ backup/muse-glimmer-30b/arxivmath/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
264
+ backup/muse-glimmer-30b/frontierscience-olympiad/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
265
+ backup/muse-glimmer-30b/frontierscience-olympiad/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
266
+ backup/muse-glimmer-30b/frontierscience-research/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
267
+ backup/muse-glimmer-30b/frontierscience-research/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
268
+ backup/muse-glimmer-30b/hle-verified/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
269
+ backup/muse-glimmer-30b/hle-verified/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
270
+ backup/muse-glimmer-30b/scicoding-agent/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
271
+ backup/muse-glimmer-30b/terminal-bench-2.1/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
272
+ backup/muse-glimmer-30b/terminal-bench-2.1/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
273
+ granite-4.2-8b/arxivmath/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
274
+ granite-4.2-8b/frontierscience-olympiad/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
275
+ granite-4.2-8b/frontierscience-olympiad/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
276
+ granite-4.2-8b/frontierscience-research/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
277
+ granite-4.2-8b/frontierscience-research/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
278
+ granite-4.2-8b/hle-verified/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
279
+ granite-4.2-8b/hle-verified/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
280
+ granite-4.2-8b/scicoding-agent/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
281
+ granite-4.2-8b/terminal-bench-2.1/judge_trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
282
+ granite-4.2-8b/terminal-bench-2.1/trajectories.jsonl filter=lfs diff=lfs merge=lfs -text
AGENTIC_CONFIG_ALIGNMENT.md CHANGED
@@ -1,3 +1,7 @@
 
 
 
 
1
  # Agentic configuration alignment audit — 2026-11-08
2
 
3
  ## Common sampling contract
 
1
+ # Publication update — 2026-11-08
2
+
3
+ SFT SciCoding has been replaced by its completed Nov 8 rerun: 35/100, all scored, budget policy `full-remaining-budget-no-tail-retry-v11`. Its TB export is unchanged while the TB rerun is ongoing. The earlier alignment table below is historical for the replaced SciCoding row. Aligned sampling does not imply all historical runs share the latest budget policy or runtime. Frozen configs accompany new Granite, Bolt, Cascade and archived Muse exports.
4
+
5
  # Agentic configuration alignment audit — 2026-11-08
6
 
7
  ## Common sampling contract
NVIDIA-Nemotron-3.5-Bolt-E4B/README.md ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # nemotron35-bolt-e4b evaluation
2
+
3
+ Published 2026-11-08.
4
+
5
+ | Benchmark | Metric | Result | Unscored |
6
+ |---|---|---:|---:|
7
+ | [arxivmath](arxivmath/) | Accuracy | 13.83% | 0 |
8
+ | [frontierscience-olympiad](frontierscience-olympiad/) | Accuracy | 13.17% | 0 |
9
+ | [frontierscience-research](frontierscience-research/) | Mean rubric | 17.72% | 0 |
10
+ | [scicoding-agent](scicoding-agent/) | avg@2 | 1.00% | 0 |
11
+ | [terminal-bench-2.1](terminal-bench-2.1/) | avg@2 | 3.37% | 1 |
12
+ | [hle-verified](hle-verified/) | Accuracy | 12.56% | 0 |
13
+
14
+ ## Protocol and artifacts
15
+
16
+ Frozen model/hardware/sampling/agent budgets are in `configs.json`; source and completeness details are in `provenance.json`. These runs use their original frozen runtime, not an inferred current default. Agentic evaluations use avg@2, retain all 100 SciCoding or 178 TB attempts in the denominator, and count native verifier success even after an abnormal agent stop. Unscored results remain null, not fabricated native zeros. Research reports normalized rubric separately from threshold accuracy. HLE Gold and Revision are combined, with subset labels and original scores retained.
17
+
18
+ Each benchmark contains trajectories.jsonl, judge_trajectories.jsonl and summary.json. Prompt benchmarks also preserve event history; agentic benchmarks preserve native verifier logs and trajectory continuation files in native_artifacts.tar.gz. See manifest.json for file sizes and SHA-256 hashes.
NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ad67893e5459caa293c22cd7a87b2cb97d2499ce511ab7d3fda4a7de3c3a5651
3
+ size 1491914
NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/judge_trajectories.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"benchmark": "arxivmath", "releases": [{"total_records": 68, "completed": 68, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.23529411764705882, "accuracy": 0.23529411764705882, "mean_latency_seconds": 29.621194251039174, "finish_reason_counts": {"stop": 68}, "length_limited_response_count": 0, "observed_max_completion_tokens": 16081, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 17, "n_trajectories_scored": 68, "reported_accuracy": 0.23529411764705882, "competition_metrics": {"arxiv/december": {"n_problems_scored": 17, "n_trajectories_scored": 68, "reported_accuracy": 0.23529411764705882}}}, "benchmark": "matharena-arxivmath-december", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "e4454fe169b62a7fd61c53174691b9dd01b452600e64835d468c926dbb0314ec", "scoring_hash": "4e9f1a9cb27f6e2579028dffa4a7085730f9d94e74454085b854411f79e377e3", "selected": 68, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 68, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 68, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.750200+00:00"}, {"total_records": 128, "completed": 128, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.109375, "accuracy": 0.109375, "mean_latency_seconds": 26.51612878436117, "finish_reason_counts": {"stop": 128}, "length_limited_response_count": 0, "observed_max_completion_tokens": 11274, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 32, "n_trajectories_scored": 128, "reported_accuracy": 0.109375, "competition_metrics": {"arxiv/february": {"n_problems_scored": 32, "n_trajectories_scored": 128, "reported_accuracy": 0.109375}}}, "benchmark": "matharena-arxivmath-february", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "52e765bc03e83c5be57f1e4af45e7ea12dc58255d1ceb3a749db9b172b7f68ca", "scoring_hash": "bb8050c752f78191d3ee258838fccf4c9c573d5a49d7f6b1922949fa5f6d55ae", "selected": 128, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 128, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 128, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.704483+00:00"}, {"total_records": 92, "completed": 92, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.1956521739130435, "accuracy": 0.1956521739130435, "mean_latency_seconds": 30.532360797453133, "finish_reason_counts": {"stop": 92}, "length_limited_response_count": 0, "observed_max_completion_tokens": 12618, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 23, "n_trajectories_scored": 92, "reported_accuracy": 0.1956521739130435, "competition_metrics": {"arxiv/january": {"n_problems_scored": 23, "n_trajectories_scored": 92, "reported_accuracy": 0.1956521739130435}}}, "benchmark": "matharena-arxivmath-january", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "d47e60cfdc70258a432843690228ef6c98eaa52c103d5ce3b7ec7ce928658585", "scoring_hash": "299568b60dfb4fcf962d429fab2f3fdbf0ccdce5972fcd36069806925cb53f5b", "selected": 92, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 92, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 92, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.704197+00:00"}, {"total_records": 124, "completed": 124, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.07258064516129033, "accuracy": 0.07258064516129033, "mean_latency_seconds": 21.13690979001788, "finish_reason_counts": {"stop": 124}, "length_limited_response_count": 0, "observed_max_completion_tokens": 10829, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 31, "n_trajectories_scored": 124, "reported_accuracy": 0.07258064516129033, "competition_metrics": {"arxiv/march": {"n_problems_scored": 31, "n_trajectories_scored": 124, "reported_accuracy": 0.07258064516129033}}}, "benchmark": "matharena-arxivmath-march", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "1e0e0ac197d35bcc7dd2b5641991588012cea1ba633b94cb51ae0be883ef3475", "scoring_hash": "30334b003a3b3c3d080ef3d17d7d186846351df351ca88615cf0c2e309c67ac9", "selected": 124, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 124, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 124, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.651175+00:00"}], "total_records": 412, "completed": 412, "accuracy": 0.1383495145631068}
NVIDIA-Nemotron-3.5-Bolt-E4B/arxivmath/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bb85994e075aa80f1a7cee4d464039de855af76cf7a562db29b6e75482ff5a2c
3
+ size 5745984
NVIDIA-Nemotron-3.5-Bolt-E4B/configs.json ADDED
The diff for this file is too large to render. See raw diff
 
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:70df2f9ed6145a900d930ac3c0476b0def31dc2637e208fed02a728d1d515f08
3
+ size 23260714
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca907512804a5925f5fb3a7918209a1c2e2bd001da37e2db3649ba32b4be9d25
3
+ size 12127622
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_records": 3000, "completed": 3000, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.13166666666666665, "accuracy": 0.13166666666666665, "mean_latency_seconds": 39.07526668899774, "finish_reason_counts": {"stop": 3000}, "length_limited_response_count": 0, "observed_max_completion_tokens": 15538, "judge_usage": {"judged_record_count": 3000, "request_count": 3001, "retry_request_count": 1, "prompt_tokens": 4104587, "cached_prompt_tokens": 2176, "cache_write_prompt_tokens": 3498828, "completion_tokens": 808843, "reasoning_tokens": 502290, "total_tokens": 4913430}, "judge_api_usage": {"request_count": 3001, "prompt_tokens": 4104587, "cached_prompt_tokens": 2176, "cache_write_prompt_tokens": 3498828, "completion_tokens": 808843, "reasoning_tokens": 502290, "total_tokens": 4913430}, "benchmark_metrics": {"track": "olympiad", "n_problems_scored": 100, "n_trajectories_scored": 3000, "aggregation": "avg@30", "reported_accuracy": 0.13166666666666665, "mean_rubric_score": null, "success_threshold": 0.5, "community_profile": null}, "benchmark": "frontierscience-olympiad", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "677a3c4243b8af0e22e2a22d7392cdbb825fcb309531a47799f4361e7d5f50aa", "scoring_hash": "5db3f7efd3c7e839c7cbb516fbd124fc207b08d82d6e533e97ae6921a6ed09c0", "selected": 3000, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 3000, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 3000, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.584581+00:00"}
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-olympiad/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2fa8a3aa430e8670da5c03431e8129104be48e80bb256a3e58456603ba8d3a86
3
+ size 54065167
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5416c3c4a9d1fd10ee10a1ae926064eaca30adad78a3c79349b4a3ec21e76dc9
3
+ size 20442794
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:40fb91c353861af33ecd4efe7e4299f4aaac8fdc3f7e1674186acb003c94bea4
3
+ size 25523828
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_records": 1800, "completed": 1800, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.17716666666666667, "accuracy": 0.006111111111111111, "mean_latency_seconds": 23.80791446976396, "finish_reason_counts": {"stop": 1800}, "length_limited_response_count": 0, "observed_max_completion_tokens": 7020, "judge_usage": {"judged_record_count": 1800, "request_count": 1824, "retry_request_count": 24, "prompt_tokens": 7032464, "cached_prompt_tokens": 82545, "cache_write_prompt_tokens": 6944447, "completion_tokens": 2455162, "reasoning_tokens": 1346395, "total_tokens": 9487626}, "judge_api_usage": {"request_count": 1824, "prompt_tokens": 7032464, "cached_prompt_tokens": 82545, "cache_write_prompt_tokens": 6944447, "completion_tokens": 2455162, "reasoning_tokens": 1346395, "total_tokens": 9487626}, "benchmark_metrics": {"track": "research", "n_problems_scored": 60, "n_trajectories_scored": 1800, "aggregation": "avg@30", "reported_accuracy": 0.006111111111111111, "mean_rubric_score": 0.17716666666666667, "success_threshold": 0.7, "community_profile": null}, "benchmark": "frontierscience-research", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "d4cddf376fd7cfc020791e8d5680c5fbf10d47758031c56654fc3144338d2591", "scoring_hash": "4fa1d4fe3cbd2a96bd7f46189b68da3a2571c42cb0590360468216a3ff685fcb", "selected": 1800, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 1800, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 1800, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.582122+00:00"}
NVIDIA-Nemotron-3.5-Bolt-E4B/frontierscience-research/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b34455a23751c71929c0d8c62f504acab11e5af5dbe6f0a8f815c83e86336f2f
3
+ size 49526379
NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3a33f4d1b2d534a8cf3b6e7e026ee50ff35de4564108b4fdb245304f13ef2ffc
3
+ size 12504773
NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01fe60f37a090d677d5b8f021157871f055d445d80a2a8474643c27bf2127e55
3
+ size 2374663
NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"benchmark": "hle-verified", "total_records": 1600, "completed": 1600, "correct": 201.0, "accuracy": 0.125625, "mean_score": 0.125625, "aggregation": "question-weighted Gold + Revision, one trial per question", "subsets": {"gold": {"total_records": 575, "completed": 575, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.06782608695652174, "accuracy": 0.06782608695652174, "mean_latency_seconds": 25.016965756528485, "finish_reason_counts": {"stop": 575}, "length_limited_response_count": 0, "observed_max_completion_tokens": 19358, "judge_usage": {"judged_record_count": 575, "request_count": 575, "retry_request_count": 0, "prompt_tokens": 492784, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 179769, "completion_tokens": 62784, "reasoning_tokens": 28518, "total_tokens": 555568}, "judge_api_usage": {"request_count": 573, "prompt_tokens": 483333, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 170324, "completion_tokens": 62636, "reasoning_tokens": 28473, "total_tokens": 545969}, "benchmark_metrics": {"accuracy": 0.06782608695652174, "calibration_error": 85.69488679722032, "calibration_minimum_n": 100}, "benchmark": "hle-verified-gold", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "d3a6c50dca68ca53c670db74dfd43e346953e70219dd8e8cb01b5f1f1f31b819", "scoring_hash": "5a8e4b649ecc8223bcffa75f89c80f94b14b5114ca036e046ca0f4af23fa867e", "selected": 575, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 575, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 575, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:11.277548+00:00"}, "revision": {"total_records": 1025, "completed": 1025, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.15804878048780488, "accuracy": 0.15804878048780488, "mean_latency_seconds": 26.03621413042553, "finish_reason_counts": {"stop": 1025}, "length_limited_response_count": 0, "observed_max_completion_tokens": 13442, "judge_usage": {"judged_record_count": 1025, "request_count": 1025, "retry_request_count": 0, "prompt_tokens": 949375, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 402790, "completion_tokens": 134950, "reasoning_tokens": 71877, "total_tokens": 1084325}, "judge_api_usage": {"request_count": 1021, "prompt_tokens": 928941, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 382368, "completion_tokens": 134589, "reasoning_tokens": 71731, "total_tokens": 1063530}, "benchmark_metrics": {"accuracy": 0.15804878048780488, "calibration_error": 84.14408838270278, "calibration_minimum_n": 100}, "benchmark": "hle-verified-revision", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "2a850b6521c575adfff585f3d468d42b1e9c422b6d5021dc809c6b08870e6b25", "scoring_hash": "fa6c9acdf75db7d26e87bb146e79a1db56a88672dc3d626a1b5c333159f0e968", "selected": 1025, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 1025, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 1025, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:11.416447+00:00"}}, "scoring": "LLM judge; original judgments retained"}
NVIDIA-Nemotron-3.5-Bolt-E4B/hle-verified/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:336564d76e8df50ee601b38088463189fa82b1f6e82d4c4f9c9ae360457389c8
3
+ size 24576533
NVIDIA-Nemotron-3.5-Bolt-E4B/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"README.md": {"size": 1434, "sha256": "5c6a0dc488482349dd104ea0e599a9915c294a50fa4c771d2fe6d721dc5b49df"}, "configs.json": {"size": 214326, "sha256": "bef27738e526b73048b4bb3a48d5d0f0401e01526ada0b82b84e803f341cfaa3"}, "provenance.json": {"size": 5160, "sha256": "181014ae829c2239ff0fe4b1a7c49b34bc295265b3b5cbcb472f4ce1d33f3ee1"}, "summary.json": {"size": 14116, "sha256": "fe533592327073e3caa581e0204abe0d2820ff54c0ce5d88798635e7716fd090"}, "arxivmath/events.jsonl.gz": {"size": 1491914, "sha256": "ad67893e5459caa293c22cd7a87b2cb97d2499ce511ab7d3fda4a7de3c3a5651"}, "arxivmath/judge_trajectories.jsonl": {"size": 193231, "sha256": "e93a3931534948434a49a34f92998200ae25d593ccc67706854240025c66b5e9"}, "arxivmath/summary.json": {"size": 6583, "sha256": "1b6b44ee1f4d113a96ebc0c86da524c0e8b146ee6322b6624b7ebc62f8466ab3"}, "arxivmath/trajectories.jsonl": {"size": 5745984, "sha256": "bb85994e075aa80f1a7cee4d464039de855af76cf7a562db29b6e75482ff5a2c"}, "frontierscience-olympiad/events.jsonl.gz": {"size": 23260714, "sha256": "70df2f9ed6145a900d930ac3c0476b0def31dc2637e208fed02a728d1d515f08"}, "frontierscience-olympiad/judge_trajectories.jsonl": {"size": 12127622, "sha256": "ca907512804a5925f5fb3a7918209a1c2e2bd001da37e2db3649ba32b4be9d25"}, "frontierscience-olympiad/summary.json": {"size": 1641, "sha256": "c0d994b3990a5ea59f6dc6e6e8ed27ea5e544534f157cdc2e07f1cc7a9ded9ff"}, "frontierscience-olympiad/trajectories.jsonl": {"size": 54065167, "sha256": "2fa8a3aa430e8670da5c03431e8129104be48e80bb256a3e58456603ba8d3a86"}, "frontierscience-research/events.jsonl.gz": {"size": 20442794, "sha256": "5416c3c4a9d1fd10ee10a1ae926064eaca30adad78a3c79349b4a3ec21e76dc9"}, "frontierscience-research/judge_trajectories.jsonl": {"size": 25523828, "sha256": "40fb91c353861af33ecd4efe7e4299f4aaac8fdc3f7e1674186acb003c94bea4"}, "frontierscience-research/summary.json": {"size": 1663, "sha256": "80e4d032c5618cdf7b8e221204fb940eb201d6f0112396a90682f4c4f197bb46"}, "frontierscience-research/trajectories.jsonl": {"size": 49526379, "sha256": "b34455a23751c71929c0d8c62f504acab11e5af5dbe6f0a8f815c83e86336f2f"}, "hle-verified/events.jsonl.gz": {"size": 12504773, "sha256": "3a33f4d1b2d534a8cf3b6e7e026ee50ff35de4564108b4fdb245304f13ef2ffc"}, "hle-verified/judge_trajectories.jsonl": {"size": 2374663, "sha256": "01fe60f37a090d677d5b8f021157871f055d445d80a2a8474643c27bf2127e55"}, "hle-verified/summary.json": {"size": 3292, "sha256": "6ec5a4e8b2631600efebd368244f755c042b7882ae95b7e26653447703ec9672"}, "hle-verified/trajectories.jsonl": {"size": 24576533, "sha256": "336564d76e8df50ee601b38088463189fa82b1f6e82d4c4f9c9ae360457389c8"}, "scicoding-agent/judge_trajectories.jsonl": {"size": 297746, "sha256": "52e31e9fd05b08034b3186f10fb15ab400407ce40791523480fa654ffcc8f164"}, "scicoding-agent/native_artifacts.tar.gz": {"size": 10497437, "sha256": "a569c7130f13882679adfac167d5ea6cc54442185e18ef5461ca087e30601c7a"}, "scicoding-agent/summary.json": {"size": 265, "sha256": "d175b5b2578838e374f3e3d94dac1cba46e898c60a13c0e61e1277554efe2252"}, "scicoding-agent/trajectories.jsonl": {"size": 11099367, "sha256": "21fa8895eb78a44433cc0e7c1b4d8fef74289957c55545ed588fcb3ca3ee2288"}, "terminal-bench-2.1/judge_trajectories.jsonl": {"size": 5943071, "sha256": "e3276954d8af800857c21031f748e05b91cada5bf86c4d8667f3db4843c1435c"}, "terminal-bench-2.1/native_artifacts.tar.gz": {"size": 629984452, "sha256": "a66a99722fa6ad6c8114903c8131778b420ad1a6b235f3f60142e2338088663a"}, "terminal-bench-2.1/summary.json": {"size": 281, "sha256": "58d37ea92ffe8e6726d311b0403721dddcda03547434bf686558b49cbdcd105d"}, "terminal-bench-2.1/trajectories.jsonl": {"size": 27499529, "sha256": "b14eaeb152b731f5bff323033119f345d55086284f6d61f5bc30b113e4530bc2"}}
NVIDIA-Nemotron-3.5-Bolt-E4B/provenance.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_run": "/workspace/data/sci-llm-training/runtime/model-eval-queue-20261105/nemotron35-bolt-e4b/run", "queue_completed": {"time": 1794086175.3629246, "result": "/workspace/data/sci-llm-training/runtime/model-eval-queue-20261105/nemotron35-bolt-e4b/run/completed.json"}, "published_at": "2026-11-08", "regenerated_for_publication": false, "rejudged_for_publication": false, "unscored_records": {"arxivmath": [], "frontierscience-olympiad": [], "frontierscience-research": [], "scicoding-agent": [], "terminal-bench-2.1": [{"key": null, "task_id": "sanitize-git-repo", "attempt_id": "sanitize-git-repo--attempt-001", "status": null, "exception_info": {"exception_type": "RuntimeError", "exception_message": "Command timed out after 120 seconds", "exception_traceback": "Traceback (most recent call last):\n File \"/usr/local/lib/python3.12/asyncio/tasks.py\", line 520, in wait_for\n return await fut\n ^^^^^^^^^\n File \"/usr/local/lib/python3.12/asyncio/subprocess.py\", line 201, in communicate\n stdin, stdout, stderr = await tasks.gather(stdin, stdout, stderr)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/usr/local/lib/python3.12/asyncio/subprocess.py\", line 181, in _read_stream\n output = await stream.read()\n ^^^^^^^^^^^^^^^^^^^\n File \"/usr/local/lib/python3.12/asyncio/streams.py\", line 706, in read\n block = await self.read(self._limit)\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/usr/local/lib/python3.12/asyncio/streams.py\", line 713, in read\n await self._wait_for_data('read')\n File \"/usr/local/lib/python3.12/asyncio/streams.py\", line 545, in _wait_for_data\n await self._waiter\nasyncio.exceptions.CancelledError\n\nThe above exception was the direct cause of the following exception:\n\nTraceback (most recent call last):\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/environments/docker/docker.py\", line 723, in _collect_buffered_output\n stdout_bytes, stderr_bytes = await asyncio.wait_for(\n ^^^^^^^^^^^^^^^^^^^^^^^\n File \"/usr/local/lib/python3.12/asyncio/tasks.py\", line 519, in wait_for\n async with timeouts.timeout(timeout):\n ^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/usr/local/lib/python3.12/asyncio/timeouts.py\", line 115, in __aexit__\n raise TimeoutError from exc_val\nTimeoutError\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/trial/trial.py\", line 415, in run\n await self._prepare()\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/trial/trial.py\", line 459, in _prepare\n await self._setup_agent()\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/trial/trial.py\", line 1574, in _setup_agent\n await asyncio.wait_for(\n File \"/usr/local/lib/python3.12/asyncio/tasks.py\", line 520, in wait_for\n return await fut\n ^^^^^^^^^\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/agents/terminus_2/terminus_2.py\", line 513, in setup\n await self._session.start()\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/agents/terminus_2/tmux_session.py\", line 498, in start\n await self._attempt_tmux_installation()\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/agents/terminus_2/tmux_session.py\", line 110, in _attempt_tmux_installation\n await asyncio.wait_for(\n File \"/usr/local/lib/python3.12/asyncio/tasks.py\", line 520, in wait_for\n return await fut\n ^^^^^^^^^\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/agents/terminus_2/tmux_session.py\", line 162, in _install_recording_tools\n result = await self.environment.exec(\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/environments/docker/docker.py\", line 1249, in exec\n return await self._compose_exec(\n ^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/environments/docker/docker.py\", line 1326, in _compose_exec\n return await self._run_docker_compose_command(\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/environments/docker/docker.py\", line 697, in _run_docker_compose_command\n result = await self._collect_buffered_output(\n ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^\n File \"/opt/scicoding/venv/lib/python3.12/site-packages/harbor/environments/docker/docker.py\", line 730, in _collect_buffered_output\n raise RuntimeError(f\"Command timed out after {timeout_sec} seconds\")\nRuntimeError: Command timed out after 120 seconds\n", "occurred_at": "2026-11-07T20:45:21.661482Z"}}], "hle-verified": []}, "comparison_target": true, "runtime_policy": "Original frozen external-model runtime, explicitly approved; not relabeled as the newer pinned runtime.", "score_policy": "Native verifier results retained independently of abnormal agent completion. No new timeout-to-zero conversions."}
NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/judge_trajectories.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/native_artifacts.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a569c7130f13882679adfac167d5ea6cc54442185e18ef5461ca087e30601c7a
3
+ size 10497437
NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"tasks": 50, "attempts": 100, "attempts_per_task": 2, "metric": "avg@2", "passed": 1, "avg_at_2": 0.01, "unscored": 0, "agent_timeouts": 0, "reward_positive_with_exception": 0, "scoring": "native_verifier", "scoring_policy": "native-verifier-after-agent-stop-v1"}
NVIDIA-Nemotron-3.5-Bolt-E4B/scicoding-agent/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:21fa8895eb78a44433cc0e7c1b4d8fef74289957c55545ed588fcb3ca3ee2288
3
+ size 11099367
NVIDIA-Nemotron-3.5-Bolt-E4B/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"model": "nemotron35-bolt-e4b", "benchmarks": {"arxivmath": {"benchmark": "arxivmath", "releases": [{"total_records": 68, "completed": 68, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.23529411764705882, "accuracy": 0.23529411764705882, "mean_latency_seconds": 29.621194251039174, "finish_reason_counts": {"stop": 68}, "length_limited_response_count": 0, "observed_max_completion_tokens": 16081, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 17, "n_trajectories_scored": 68, "reported_accuracy": 0.23529411764705882, "competition_metrics": {"arxiv/december": {"n_problems_scored": 17, "n_trajectories_scored": 68, "reported_accuracy": 0.23529411764705882}}}, "benchmark": "matharena-arxivmath-december", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "e4454fe169b62a7fd61c53174691b9dd01b452600e64835d468c926dbb0314ec", "scoring_hash": "4e9f1a9cb27f6e2579028dffa4a7085730f9d94e74454085b854411f79e377e3", "selected": 68, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 68, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 68, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.750200+00:00"}, {"total_records": 128, "completed": 128, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.109375, "accuracy": 0.109375, "mean_latency_seconds": 26.51612878436117, "finish_reason_counts": {"stop": 128}, "length_limited_response_count": 0, "observed_max_completion_tokens": 11274, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 32, "n_trajectories_scored": 128, "reported_accuracy": 0.109375, "competition_metrics": {"arxiv/february": {"n_problems_scored": 32, "n_trajectories_scored": 128, "reported_accuracy": 0.109375}}}, "benchmark": "matharena-arxivmath-february", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "52e765bc03e83c5be57f1e4af45e7ea12dc58255d1ceb3a749db9b172b7f68ca", "scoring_hash": "bb8050c752f78191d3ee258838fccf4c9c573d5a49d7f6b1922949fa5f6d55ae", "selected": 128, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 128, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 128, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.704483+00:00"}, {"total_records": 92, "completed": 92, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.1956521739130435, "accuracy": 0.1956521739130435, "mean_latency_seconds": 30.532360797453133, "finish_reason_counts": {"stop": 92}, "length_limited_response_count": 0, "observed_max_completion_tokens": 12618, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 23, "n_trajectories_scored": 92, "reported_accuracy": 0.1956521739130435, "competition_metrics": {"arxiv/january": {"n_problems_scored": 23, "n_trajectories_scored": 92, "reported_accuracy": 0.1956521739130435}}}, "benchmark": "matharena-arxivmath-january", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "d47e60cfdc70258a432843690228ef6c98eaa52c103d5ce3b7ec7ce928658585", "scoring_hash": "299568b60dfb4fcf962d429fab2f3fdbf0ccdce5972fcd36069806925cb53f5b", "selected": 92, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 92, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 92, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.704197+00:00"}, {"total_records": 124, "completed": 124, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.07258064516129033, "accuracy": 0.07258064516129033, "mean_latency_seconds": 21.13690979001788, "finish_reason_counts": {"stop": 124}, "length_limited_response_count": 0, "observed_max_completion_tokens": 10829, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 31, "n_trajectories_scored": 124, "reported_accuracy": 0.07258064516129033, "competition_metrics": {"arxiv/march": {"n_problems_scored": 31, "n_trajectories_scored": 124, "reported_accuracy": 0.07258064516129033}}}, "benchmark": "matharena-arxivmath-march", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "1e0e0ac197d35bcc7dd2b5641991588012cea1ba633b94cb51ae0be883ef3475", "scoring_hash": "30334b003a3b3c3d080ef3d17d7d186846351df351ca88615cf0c2e309c67ac9", "selected": 124, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 124, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 124, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.651175+00:00"}], "total_records": 412, "completed": 412, "accuracy": 0.1383495145631068}, "frontierscience-olympiad": {"total_records": 3000, "completed": 3000, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.13166666666666665, "accuracy": 0.13166666666666665, "mean_latency_seconds": 39.07526668899774, "finish_reason_counts": {"stop": 3000}, "length_limited_response_count": 0, "observed_max_completion_tokens": 15538, "judge_usage": {"judged_record_count": 3000, "request_count": 3001, "retry_request_count": 1, "prompt_tokens": 4104587, "cached_prompt_tokens": 2176, "cache_write_prompt_tokens": 3498828, "completion_tokens": 808843, "reasoning_tokens": 502290, "total_tokens": 4913430}, "judge_api_usage": {"request_count": 3001, "prompt_tokens": 4104587, "cached_prompt_tokens": 2176, "cache_write_prompt_tokens": 3498828, "completion_tokens": 808843, "reasoning_tokens": 502290, "total_tokens": 4913430}, "benchmark_metrics": {"track": "olympiad", "n_problems_scored": 100, "n_trajectories_scored": 3000, "aggregation": "avg@30", "reported_accuracy": 0.13166666666666665, "mean_rubric_score": null, "success_threshold": 0.5, "community_profile": null}, "benchmark": "frontierscience-olympiad", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "677a3c4243b8af0e22e2a22d7392cdbb825fcb309531a47799f4361e7d5f50aa", "scoring_hash": "5db3f7efd3c7e839c7cbb516fbd124fc207b08d82d6e533e97ae6921a6ed09c0", "selected": 3000, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 3000, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 3000, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.584581+00:00"}, "frontierscience-research": {"total_records": 1800, "completed": 1800, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.17716666666666667, "accuracy": 0.006111111111111111, "mean_latency_seconds": 23.80791446976396, "finish_reason_counts": {"stop": 1800}, "length_limited_response_count": 0, "observed_max_completion_tokens": 7020, "judge_usage": {"judged_record_count": 1800, "request_count": 1824, "retry_request_count": 24, "prompt_tokens": 7032464, "cached_prompt_tokens": 82545, "cache_write_prompt_tokens": 6944447, "completion_tokens": 2455162, "reasoning_tokens": 1346395, "total_tokens": 9487626}, "judge_api_usage": {"request_count": 1824, "prompt_tokens": 7032464, "cached_prompt_tokens": 82545, "cache_write_prompt_tokens": 6944447, "completion_tokens": 2455162, "reasoning_tokens": 1346395, "total_tokens": 9487626}, "benchmark_metrics": {"track": "research", "n_problems_scored": 60, "n_trajectories_scored": 1800, "aggregation": "avg@30", "reported_accuracy": 0.006111111111111111, "mean_rubric_score": 0.17716666666666667, "success_threshold": 0.7, "community_profile": null}, "benchmark": "frontierscience-research", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "d4cddf376fd7cfc020791e8d5680c5fbf10d47758031c56654fc3144338d2591", "scoring_hash": "4fa1d4fe3cbd2a96bd7f46189b68da3a2571c42cb0590360468216a3ff685fcb", "selected": 1800, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 1800, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 1800, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:13.582122+00:00"}, "scicoding-agent": {"tasks": 50, "attempts": 100, "attempts_per_task": 2, "metric": "avg@2", "passed": 1, "avg_at_2": 0.01, "unscored": 0, "agent_timeouts": 0, "reward_positive_with_exception": 0, "scoring": "native_verifier", "scoring_policy": "native-verifier-after-agent-stop-v1"}, "terminal-bench-2.1": {"tasks": 89, "attempts": 178, "attempts_per_task": 2, "metric": "avg@2", "passed": 6, "avg_at_2": 0.033707865168539325, "unscored": 1, "agent_timeouts": 4, "reward_positive_with_exception": 0, "scoring": "native_verifier", "scoring_policy": "native-verifier-after-agent-stop-v1"}, "hle-verified": {"benchmark": "hle-verified", "total_records": 1600, "completed": 1600, "correct": 201.0, "accuracy": 0.125625, "mean_score": 0.125625, "aggregation": "question-weighted Gold + Revision, one trial per question", "subsets": {"gold": {"total_records": 575, "completed": 575, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.06782608695652174, "accuracy": 0.06782608695652174, "mean_latency_seconds": 25.016965756528485, "finish_reason_counts": {"stop": 575}, "length_limited_response_count": 0, "observed_max_completion_tokens": 19358, "judge_usage": {"judged_record_count": 575, "request_count": 575, "retry_request_count": 0, "prompt_tokens": 492784, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 179769, "completion_tokens": 62784, "reasoning_tokens": 28518, "total_tokens": 555568}, "judge_api_usage": {"request_count": 573, "prompt_tokens": 483333, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 170324, "completion_tokens": 62636, "reasoning_tokens": 28473, "total_tokens": 545969}, "benchmark_metrics": {"accuracy": 0.06782608695652174, "calibration_error": 85.69488679722032, "calibration_minimum_n": 100}, "benchmark": "hle-verified-gold", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "d3a6c50dca68ca53c670db74dfd43e346953e70219dd8e8cb01b5f1f1f31b819", "scoring_hash": "5a8e4b649ecc8223bcffa75f89c80f94b14b5114ca036e046ca0f4af23fa867e", "selected": 575, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 575, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 575, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:11.277548+00:00"}, "revision": {"total_records": 1025, "completed": 1025, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.15804878048780488, "accuracy": 0.15804878048780488, "mean_latency_seconds": 26.03621413042553, "finish_reason_counts": {"stop": 1025}, "length_limited_response_count": 0, "observed_max_completion_tokens": 13442, "judge_usage": {"judged_record_count": 1025, "request_count": 1025, "retry_request_count": 0, "prompt_tokens": 949375, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 402790, "completion_tokens": 134950, "reasoning_tokens": 71877, "total_tokens": 1084325}, "judge_api_usage": {"request_count": 1021, "prompt_tokens": 928941, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 382368, "completion_tokens": 134589, "reasoning_tokens": 71731, "total_tokens": 1063530}, "benchmark_metrics": {"accuracy": 0.15804878048780488, "calibration_error": 84.14408838270278, "calibration_minimum_n": 100}, "benchmark": "hle-verified-revision", "model": "nemotron35-bolt-e4b", "execution_mode": "non_agent", "generation_hash": "2a850b6521c575adfff585f3d468d42b1e9c422b6d5021dc809c6b08870e6b25", "scoring_hash": "fa6c9acdf75db7d26e87bb146e79a1db56a88672dc3d626a1b5c333159f0e968", "selected": 1025, "incomplete": 0, "generated_now": 0, "scored_now": 0, "cached_completed": 1025, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 1025, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-07T19:48:11.416447+00:00"}}, "scoring": "LLM judge; original judgments retained"}}, "generation_finished": true, "fully_scored": false, "unscored": {"arxivmath": 0, "frontierscience-olympiad": 0, "frontierscience-research": 0, "scicoding-agent": 0, "terminal-bench-2.1": 1, "hle-verified": 0}}
NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e3276954d8af800857c21031f748e05b91cada5bf86c4d8667f3db4843c1435c
3
+ size 5943071
NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/native_artifacts.tar.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a66a99722fa6ad6c8114903c8131778b420ad1a6b235f3f60142e2338088663a
3
+ size 629984452
NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"tasks": 89, "attempts": 178, "attempts_per_task": 2, "metric": "avg@2", "passed": 6, "avg_at_2": 0.033707865168539325, "unscored": 1, "agent_timeouts": 4, "reward_positive_with_exception": 0, "scoring": "native_verifier", "scoring_policy": "native-verifier-after-agent-stop-v1"}
NVIDIA-Nemotron-3.5-Bolt-E4B/terminal-bench-2.1/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b14eaeb152b731f5bff323033119f345d55086284f6d61f5bc30b113e4530bc2
3
+ size 27499529
Nemotron-Cascade-2-30B-A3B/README.md ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # nemotron-cascade2-30b evaluation
2
+
3
+ Published 2026-11-08.
4
+
5
+ | Benchmark | Metric | Result | Unscored |
6
+ |---|---|---:|---:|
7
+ | [arxivmath](arxivmath/) | Accuracy | 29.13% | 0 |
8
+ | [frontierscience-olympiad](frontierscience-olympiad/) | Accuracy | 47.90% | 0 |
9
+ | [frontierscience-research](frontierscience-research/) | Mean rubric | 30.84% | 0 |
10
+ | [scicoding-agent](scicoding-agent/) | avg@2 | 33.00% | 0 |
11
+ | [terminal-bench-2.1](terminal-bench-2.1/) | avg@2 | 25.28% | 2 |
12
+ | [hle-verified](hle-verified/) | Accuracy | 26.81% | 0 |
13
+
14
+ ## Protocol and artifacts
15
+
16
+ Frozen model/hardware/sampling/agent budgets are in `configs.json`; source and completeness details are in `provenance.json`. These runs use their original frozen runtime, not an inferred current default. Agentic evaluations use avg@2, retain all 100 SciCoding or 178 TB attempts in the denominator, and count native verifier success even after an abnormal agent stop. Unscored results remain null, not fabricated native zeros. Research reports normalized rubric separately from threshold accuracy. HLE Gold and Revision are combined, with subset labels and original scores retained.
17
+
18
+ Each benchmark contains trajectories.jsonl, judge_trajectories.jsonl and summary.json. Prompt benchmarks also preserve event history; agentic benchmarks preserve native verifier logs and trajectory continuation files in native_artifacts.tar.gz. See manifest.json for file sizes and SHA-256 hashes.
Nemotron-Cascade-2-30B-A3B/arxivmath/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b5edd885b155590b2f50ba390b4ff4f6b3b35c56a1bb92f5f5b629c14cd3fad6
3
+ size 52950562
Nemotron-Cascade-2-30B-A3B/arxivmath/judge_trajectories.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
Nemotron-Cascade-2-30B-A3B/arxivmath/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"benchmark": "arxivmath", "releases": [{"total_records": 68, "completed": 68, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.3235294117647059, "accuracy": 0.3235294117647059, "mean_latency_seconds": 870.663082991903, "finish_reason_counts": {"length": 21, "stop": 47}, "length_limited_response_count": 21, "observed_max_completion_tokens": 98304, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 17, "n_trajectories_scored": 68, "reported_accuracy": 0.3235294117647059, "competition_metrics": {"arxiv/december": {"n_problems_scored": 17, "n_trajectories_scored": 68, "reported_accuracy": 0.3235294117647059}}}, "benchmark": "matharena-arxivmath-december", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "6e59e4fe6739578b290732824326aa322039ac5dca81e13de48219d76c47f1f7", "scoring_hash": "cb6b61d729f910744ef658c17fda917eb794cf0b947d01f67161a8e17666cae9", "selected": 68, "incomplete": 0, "generated_now": 68, "scored_now": 68, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 68, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T01:07:16.683138+00:00"}, {"total_records": 128, "completed": 128, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.203125, "accuracy": 0.203125, "mean_latency_seconds": 732.3832417432277, "finish_reason_counts": {"length": 40, "stop": 88}, "length_limited_response_count": 40, "observed_max_completion_tokens": 98304, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 32, "n_trajectories_scored": 128, "reported_accuracy": 0.203125, "competition_metrics": {"arxiv/february": {"n_problems_scored": 32, "n_trajectories_scored": 128, "reported_accuracy": 0.203125}}}, "benchmark": "matharena-arxivmath-february", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "96c69c8070a32965d0ef607be371cd6e31836caf5930c01cd226c0e7930cd4db", "scoring_hash": "6acd96fe2d22bb9a28279d7d8674c6cdf856fcd50d94474b6765afe551400bda", "selected": 128, "incomplete": 0, "generated_now": 128, "scored_now": 128, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 128, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T01:19:00.961708+00:00"}, {"total_records": 92, "completed": 92, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.391304347826087, "accuracy": 0.391304347826087, "mean_latency_seconds": 888.3882868897598, "finish_reason_counts": {"length": 26, "stop": 66}, "length_limited_response_count": 26, "observed_max_completion_tokens": 98304, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 23, "n_trajectories_scored": 92, "reported_accuracy": 0.391304347826087, "competition_metrics": {"arxiv/january": {"n_problems_scored": 23, "n_trajectories_scored": 92, "reported_accuracy": 0.391304347826087}}}, "benchmark": "matharena-arxivmath-january", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "2d8b032698c5b6006bcd16330f4a75fa4035ccf4934929ac887515b8aa2a3623", "scoring_hash": "1a95a3120e03eea7c523a8e9bc0c8a806d1a3cb34c0da6694ba959baddd15dce", "selected": 92, "incomplete": 0, "generated_now": 92, "scored_now": 92, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 92, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T01:15:21.868239+00:00"}, {"total_records": 124, "completed": 124, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.2903225806451613, "accuracy": 0.2903225806451613, "mean_latency_seconds": 700.6650238223618, "finish_reason_counts": {"length": 10, "stop": 114}, "length_limited_response_count": 10, "observed_max_completion_tokens": 98304, "judge_usage": {"judged_record_count": 0, "request_count": 0, "retry_request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "judge_api_usage": {"request_count": 0, "prompt_tokens": 0, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 0, "completion_tokens": 0, "reasoning_tokens": 0, "total_tokens": 0}, "benchmark_metrics": {"competition": "arxivmath", "aggregation": "avg@4", "n_problems_scored": 31, "n_trajectories_scored": 124, "reported_accuracy": 0.2903225806451613, "competition_metrics": {"arxiv/march": {"n_problems_scored": 31, "n_trajectories_scored": 124, "reported_accuracy": 0.2903225806451613}}}, "benchmark": "matharena-arxivmath-march", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "2c7833efff99cfded36b0148298d18be16ffb87ea173d8e9d3416880d40db8c1", "scoring_hash": "b8df5ba22e02c54a6e338830a8977039ddb0177ec7782a3d4775fcc808106070", "selected": 124, "incomplete": 0, "generated_now": 124, "scored_now": 124, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 124, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T01:08:53.150078+00:00"}], "total_records": 412, "completed": 412, "accuracy": 0.2912621359223301}
Nemotron-Cascade-2-30B-A3B/arxivmath/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f2fbe7b11fe8363ef7671c3f8e4bc74a541fa60689864f2b5df87dcc20ada3b8
3
+ size 97867871
Nemotron-Cascade-2-30B-A3B/configs.json ADDED
The diff for this file is too large to render. See raw diff
 
Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5fe279e89bbf9920b38c29a8646b8b6c8293cd1f8e0b50213968cf0730fbbfbf
3
+ size 64972272
Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8cd22585233705379b04ca279bfe1372e77707a5b802427b063cee763ece6ab2
3
+ size 10716324
Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_records": 3000, "completed": 3000, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.479, "accuracy": 0.479, "mean_latency_seconds": 113.09971219053853, "finish_reason_counts": {"stop": 3000}, "length_limited_response_count": 0, "observed_max_completion_tokens": 91122, "judge_usage": {"judged_record_count": 3000, "request_count": 3000, "retry_request_count": 0, "prompt_tokens": 2666162, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 1286295, "completion_tokens": 776774, "reasoning_tokens": 542844, "total_tokens": 3442936}, "judge_api_usage": {"request_count": 3000, "prompt_tokens": 2666162, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 1286295, "completion_tokens": 776774, "reasoning_tokens": 542844, "total_tokens": 3442936}, "benchmark_metrics": {"track": "olympiad", "n_problems_scored": 100, "n_trajectories_scored": 3000, "aggregation": "avg@30", "reported_accuracy": 0.479, "mean_rubric_score": null, "success_threshold": 0.5, "community_profile": null}, "benchmark": "frontierscience-olympiad", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "fe0328ec10ca5c843f9538d25a0e27aa8953a86ca83b80bbd92860321f388292", "scoring_hash": "0e899897c7cf45f80e15edc01a869181d6f588ca19d5dc3bcb855d8325f66209", "selected": 3000, "incomplete": 0, "generated_now": 3000, "scored_now": 3000, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 3000, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T00:58:45.805857+00:00"}
Nemotron-Cascade-2-30B-A3B/frontierscience-olympiad/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ea3498ddd70a6769487fca7fe5f8a13b71388983161db7968bee0c2b8b9f2abf
3
+ size 117733436
Nemotron-Cascade-2-30B-A3B/frontierscience-research/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b749d33b5159d1f92b153368e564b89cbe1a18266d4b017e0168bf52c41cc3ea
3
+ size 59094537
Nemotron-Cascade-2-30B-A3B/frontierscience-research/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:beab025116069ab5e4857f9a940757adb747833ddee31335e73ff8228c78c655
3
+ size 28526852
Nemotron-Cascade-2-30B-A3B/frontierscience-research/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_records": 1800, "completed": 1800, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.3084307407407407, "accuracy": 0.06944444444444445, "mean_latency_seconds": 102.03766752248546, "finish_reason_counts": {"stop": 1800}, "length_limited_response_count": 0, "observed_max_completion_tokens": 84335, "judge_usage": {"judged_record_count": 1800, "request_count": 1832, "retry_request_count": 32, "prompt_tokens": 14601052, "cached_prompt_tokens": 243632, "cache_write_prompt_tokens": 14351073, "completion_tokens": 3273395, "reasoning_tokens": 1993080, "total_tokens": 17874447}, "judge_api_usage": {"request_count": 1832, "prompt_tokens": 14601052, "cached_prompt_tokens": 243632, "cache_write_prompt_tokens": 14351073, "completion_tokens": 3273395, "reasoning_tokens": 1993080, "total_tokens": 17874447}, "benchmark_metrics": {"track": "research", "n_problems_scored": 60, "n_trajectories_scored": 1800, "aggregation": "avg@30", "reported_accuracy": 0.06944444444444445, "mean_rubric_score": 0.3084307407407408, "success_threshold": 0.7, "community_profile": null}, "benchmark": "frontierscience-research", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "91861839ae6ae8f82119246f7c98623038f186d29f71c987941fbfe3ee7b979f", "scoring_hash": "e94267060b6ed01e455b1c25c69625896aaa44778fa60f6eddc9d577e40412b5", "selected": 1800, "incomplete": 0, "generated_now": 1800, "scored_now": 1800, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 1800, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T00:34:55.768431+00:00"}
Nemotron-Cascade-2-30B-A3B/frontierscience-research/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc66990a4165d9af20d4e9e25027346382b5a73933b01279fc82ae1fe373e0cb
3
+ size 116530528
Nemotron-Cascade-2-30B-A3B/hle-verified/events.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:26a614947eadd5a10d8ca284a58b9d6b38aef822eab7e8f2470ab9c4da9df0a4
3
+ size 64297554
Nemotron-Cascade-2-30B-A3B/hle-verified/judge_trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ab325d0b99da7be84c4f17ad88a4e16c7bc1c1858e62c8d93e930535da9461f3
3
+ size 2391336
Nemotron-Cascade-2-30B-A3B/hle-verified/summary.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"benchmark": "hle-verified", "total_records": 1600, "completed": 1600, "correct": 429.0, "accuracy": 0.268125, "mean_score": 0.268125, "aggregation": "question-weighted Gold + Revision, one trial per question", "subsets": {"gold": {"total_records": 575, "completed": 575, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.24695652173913044, "accuracy": 0.24695652173913044, "mean_latency_seconds": 226.31174250935163, "finish_reason_counts": {"length": 4, "stop": 571}, "length_limited_response_count": 4, "observed_max_completion_tokens": 98304, "judge_usage": {"judged_record_count": 575, "request_count": 575, "retry_request_count": 0, "prompt_tokens": 622406, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 393616, "completion_tokens": 84230, "reasoning_tokens": 49575, "total_tokens": 706636}, "judge_api_usage": {"request_count": 575, "prompt_tokens": 622406, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 393616, "completion_tokens": 84230, "reasoning_tokens": 49575, "total_tokens": 706636}, "benchmark_metrics": {"accuracy": 0.24695652173913044, "calibration_error": 65.80430438324167, "calibration_minimum_n": 100}, "benchmark": "hle-verified-gold", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "762b19300701594af3adaef4ffca265785528b3f137f4cb0f1f157080947b125", "scoring_hash": "895570e4253643f5357012b769acfe2f84195940115dc89f520077247288df07", "selected": 575, "incomplete": 0, "generated_now": 575, "scored_now": 575, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 575, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T00:19:06.887911+00:00"}, "revision": {"total_records": 1025, "completed": 1025, "query_errors": 0, "score_errors": 0, "generated_unscored": 0, "mean_score": 0.28, "accuracy": 0.28, "mean_latency_seconds": 205.19389040656569, "finish_reason_counts": {"length": 8, "stop": 1017}, "length_limited_response_count": 8, "observed_max_completion_tokens": 98304, "judge_usage": {"judged_record_count": 1025, "request_count": 1025, "retry_request_count": 0, "prompt_tokens": 1204110, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 839150, "completion_tokens": 161510, "reasoning_tokens": 96391, "total_tokens": 1365620}, "judge_api_usage": {"request_count": 1025, "prompt_tokens": 1204110, "cached_prompt_tokens": 0, "cache_write_prompt_tokens": 839150, "completion_tokens": 161510, "reasoning_tokens": 96391, "total_tokens": 1365620}, "benchmark_metrics": {"accuracy": 0.28, "calibration_error": 64.47774363600391, "calibration_minimum_n": 100}, "benchmark": "hle-verified-revision", "model": "nemotron-cascade2-30b", "execution_mode": "non_agent", "generation_hash": "d142f1acd3773f86b0ab44bdde2a61c05ba8730359756024caf815cedfa11512", "scoring_hash": "2fff682c7e9d0bd2ed244126c2a95cb5d582b848a6131b0e1a7dc06e2dddcb2f", "selected": 1025, "incomplete": 0, "generated_now": 1025, "scored_now": 1025, "cached_completed": 0, "stage": "all", "worker_count": 1, "worker_index": 0, "worker_owned_predictions": 1025, "generation_concurrency": 256, "judge_concurrency": 256, "updated_at": "2026-11-08T00:27:16.730713+00:00"}}, "scoring": "LLM judge; original judgments retained"}
Nemotron-Cascade-2-30B-A3B/hle-verified/trajectories.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d092e03f303ad1b112fb87ee6cd940c9adc0bffaeb914139cd7c27fe2dc02c78
3
+ size 115169009
Nemotron-Cascade-2-30B-A3B/manifest.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"README.md": {"size": 1438, "sha256": "b3f7906b2b7480716f96890b375ba0061d466417cbdfb5ef73b487edb550bb49"}, "configs.json": {"size": 214397, "sha256": "61b510fa5e5531688494d7b0d8b86616463c6a21bbb15c984c149ee05494799a"}, "provenance.json": {"size": 17275, "sha256": "f16ed8a16a56d849d93a61975e69bbf49f6d36dfd1354963ca75fab5f3322225"}, "summary.json": {"size": 14140, "sha256": "8608b39f83d84e5a49713078aeec75d076c17e71383c43746146aa724ff9dd4d"}, "arxivmath/events.jsonl.gz": {"size": 52950562, "sha256": "b5edd885b155590b2f50ba390b4ff4f6b3b35c56a1bb92f5f5b629c14cd3fad6"}, "arxivmath/judge_trajectories.jsonl": {"size": 194241, "sha256": "b9ffa46d10a084973d322024b7a11621eaddda394e11f9d0873474577e912059"}, "arxivmath/summary.json": {"size": 6641, "sha256": "68f82f12a5e2b9d6e25085d42ec68abe318ac7aba2ab0c9535498b1c8decad2d"}, "arxivmath/trajectories.jsonl": {"size": 97867871, "sha256": "f2fbe7b11fe8363ef7671c3f8e4bc74a541fa60689864f2b5df87dcc20ada3b8"}, "frontierscience-olympiad/events.jsonl.gz": {"size": 64972272, "sha256": "5fe279e89bbf9920b38c29a8646b8b6c8293cd1f8e0b50213968cf0730fbbfbf"}, "frontierscience-olympiad/judge_trajectories.jsonl": {"size": 10716324, "sha256": "8cd22585233705379b04ca279bfe1372e77707a5b802427b063cee763ece6ab2"}, "frontierscience-olympiad/summary.json": {"size": 1599, "sha256": "55e6eb1650808d52b9c5f123469d041e09984a1f52e8640e7f16bedfa841c1fd"}, "frontierscience-olympiad/trajectories.jsonl": {"size": 117733436, "sha256": "ea3498ddd70a6769487fca7fe5f8a13b71388983161db7968bee0c2b8b9f2abf"}, "frontierscience-research/events.jsonl.gz": {"size": 59094537, "sha256": "b749d33b5159d1f92b153368e564b89cbe1a18266d4b017e0168bf52c41cc3ea"}, "frontierscience-research/judge_trajectories.jsonl": {"size": 28526852, "sha256": "beab025116069ab5e4857f9a940757adb747833ddee31335e73ff8228c78c655"}, "frontierscience-research/summary.json": {"size": 1674, "sha256": "5263b0a612c4a714b8aab8acad710bafc5b78bf453d5c13f9ba73222276d1df8"}, "frontierscience-research/trajectories.jsonl": {"size": 116530528, "sha256": "fc66990a4165d9af20d4e9e25027346382b5a73933b01279fc82ae1fe373e0cb"}, "hle-verified/events.jsonl.gz": {"size": 64297554, "sha256": "26a614947eadd5a10d8ca284a58b9d6b38aef822eab7e8f2470ab9c4da9df0a4"}, "hle-verified/judge_trajectories.jsonl": {"size": 2391336, "sha256": "ab325d0b99da7be84c4f17ad88a4e16c7bc1c1858e62c8d93e930535da9461f3"}, "hle-verified/summary.json": {"size": 3285, "sha256": "6606480269cccbb6c0027d6ce3766d865c6c4f68bbc6412d3b11883fd2cfdc7f"}, "hle-verified/trajectories.jsonl": {"size": 115169009, "sha256": "d092e03f303ad1b112fb87ee6cd940c9adc0bffaeb914139cd7c27fe2dc02c78"}, "scicoding-agent/judge_trajectories.jsonl": {"size": 352540, "sha256": "710f65988af3ed176ff397cda2d6b21c7f849fee700fd95476158708ed11b125"}, "scicoding-agent/native_artifacts.tar.gz": {"size": 86662605, "sha256": "5906c4e97481838093f1aac633d2278971da7aa72b33b077d5b1d8a603e811d0"}, "scicoding-agent/summary.json": {"size": 266, "sha256": "f3c81e6828113e04849d000259e7c4f4c729ac4785bf93f35b25f15ac2cbd5a9"}, "scicoding-agent/trajectories.jsonl": {"size": 46525937, "sha256": "785f8e032fba51fbc284b4981d8cf1f0fb341a8cd6c35b83673e78dd5e5d8806"}, "terminal-bench-2.1/judge_trajectories.jsonl": {"size": 4315869, "sha256": "d9158047780ccf29bb621f3ee35f212900714529320a863d53857cba4cbc1bbb"}, "terminal-bench-2.1/native_artifacts.tar.gz": {"size": 504918238, "sha256": "695cc6187441b7b6a45a034a323958642492db0a28fa2cce3d65ffec77b3bd8f"}, "terminal-bench-2.1/summary.json": {"size": 282, "sha256": "60a6235075a4bf9c805f86412a592328c442b5e9cfb623e0b25bef375fb9215a"}, "terminal-bench-2.1/trajectories.jsonl": {"size": 96067689, "sha256": "2c57632ebcc904e8ea2421427dbdf73810457480ec78e5b98f26726af9506e1d"}}