nodejs/node

2,504,694events in the training captures

7models serving questions

2020 → 2026span of the captures

22results

oss · nodejs__node · World model · public

01 · Data

Dataset versions

A dataset version is a capture of the source as of a date: its event spine and the files training reads.

5 dataset versions

nodejs/node nodejs__node @ asof-2026-09-01

pooled8's and the serving model's latest zero-shot export target; recorded where it is (the archive drive, _vei_out/envs/node-asof-2026-09-01/nodejs__node/events); see REPORT.md

llm_report · REPORT.md

nodejs__node@asof-2026-09-01 — data stage report

Recorded 2026-10-13 as part of the OSS line's migration to the factory. The data already existed and nothing was rematerialized.

Inputs, by identity. GitHub's event API for nodejs/node, materialized into the OSS canonical spine. Pooled8's and the serving model's latest zero-shot export target. The raw captures stay external, on the archive drive and Modal.

Output, by identity. canonical_events.jsonl, 839,725 events from 2020-01-01T15:11:11Z to 2026-09-01T13:52:47Z. The spine is recorded where it already is (the archive drive, _vei_out/envs/node-asof-2026-09-01/nodejs__node/events) and not copied into the registry.

Drift. Ingest-time check (drift.json, drift.txt): 839,725 events over 81 periods | structural types excluded: none; 27 verb flags, 4 volume flags. The OSS runs split each stream at its own time quantile, so no single calendar split applies across the pooled streams, and no split-time check was run here. Its absence is a caveat on any result.

Receipts (3)

drift.json

{
  "era": {
    "n_periods": 81,
    "n_events": 839725,
    "structural_types": [],
    "verb_flags": [
      {
        "period": "2026-07",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 748,
        "trailing_median": 95.0,
        "ratio": 7.873684210526315
      },
      {
        "period": "2023-09",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 139,
        "trailing_median": 20.0,
        "ratio": 6.95
      },
      {
        "period": "2026-08",
        "verb": "pr_closed",
        "kind": "spike",
        "count": 745,
        "trailing_median": 125.5,
        "ratio": 5.936254980079681
      },
      {
        "period": "2026-05",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 342,
        "trailing_median": 80.5,
        "ratio": 4.248447204968944
      },
      {
        "period": "2020-09",
        "verb": "unlabeled",
        "kind": "spike",
        "count": 401,
        "trailing_median": 96.5,
        "ratio": 4.155440414507772
      },
      {
        "period": "2026-09",
        "verb": "renamed",
        "kind": "dropout",
        "count": 0,
        "trailing_median": 51.5,
        "ratio": 0.0
      },
      {
        "period": "2026-09",
        "verb": "issue_closed",
        "kind": "dropout",
        "count": 2,
        "trailing_median": 306.5,
        "ratio": 0.0065252854812398045
      },
      {
        "period": "2026-09",
        "verb": "cross_referenced",
        "kind": "dropout",
        "count": 11,
        "trailing_median": 1676.5,
        "ratio": 0.00656128839844915
      },
      {
        "period": "2026-09",
        "verb": "review_inline_comment",
        "kind": "dropout",
        "count": 10,
        "trailing_median": 747.0,
        "ratio": 0.013386880856760375
      },
      {
        "period": "2026-09",
        "verb": "review_commented",
        "kind": "dropout",
        "count": 9,
        "trailing_median": 624.0,
        "ratio": 0.014423076923076924
      },
      {
        "period": "2026-09",
        "verb": "pr_commit",
        "kind": "dropout",
        "count": 44,
        "trailing_median": 2030.0,
        "ratio": 0.02167487684729064
      },
      {
        "period": "2026-09",
        "verb": "pr_closed",
        "kind": "dropout",
        "count": 3,
        "trailing_median": 136.5,
        "ratio": 0.02197802197802198
      },
      {
        "period": "2026-09",
        "verb": "review_approved",
        "kind": "dropout",
        "count": 22,
        "trailing_median": 947.0,
        "ratio": 0.023231256599788808
      },
      {
        "period": "2026-09",
        "verb": "issue_comment",
        "kind": "dropout",
        "count": 23,
        "trailing_median": 932.0,
        "ratio": 0.02467811158798283
      },
      {
        "period": "2026-09",
        "verb": "labeled",
        "kind": "dropout",
        "count": 56,
        "trailing_median": 2030.0,
        "ratio": 0.027586206896551724
      },
      {
        "period": "2026-09",
        "verb": "pr_opened",
        "kind": "dropout",
        "count": 12,
        "trailing_median": 433.5,
        "ratio": 0.02768166089965398
      },
      {
        "period": "2026-09",
        "verb": "head_force_pushed",
        "kind": "dropout",
        "count": 14,
        "trailing_median": 496.5,
        "ratio": 0.028197381671701913
      },
      {
        "period": "2026-09",
        "verb": "pr_comment",
        "kind": "dropout",
        "count": 77,
        "trailing_median": 2595.0,
        "ratio": 0.029672447013487476
      },
      {
        "period": "2026-09",
        "verb": "pr_merged",
        "kind": "dropout",
        "count": 7,
        "trailing_median": 233.0,
        "ratio": 0.030042918454935622
      },
      {
        "period": "2026-09",
        "verb": "issue_opened",
        "kind": "dropout",
        "count": 4,
        "trailing_median": 132.0,
        "ratio": 0.030303030303030304
      },
      {
        "period": "2026-09",
        "verb": "review_changes_requested",
        "kind": "dropout",
        "count": 1,
        "trailing_median": 33.0,
        "ratio": 0.030303030303030304
      },
      {
        "period": "2026-09",
        "verb": "ready_for_review",
        "kind": "dropout",
        "count": 2,
        "trailing_median": 42.5,
        "ratio": 0.047058823529411764
      },
      {
        "period": "2026-09",
        "verb": "unlabeled",
        "kind": "dropout",
        "count": 29,
        "trailing_median": 613.5,
        "ratio": 0.04726976365118175
      },
      {
        "period": "2026-09",
        "verb": "review_requested",
        "kind": "dropout",
        "count": 10,
        "trailing_median": 199.5,
        "ratio": 0.05012531328320802
      },
      {
        "period": "2021-02",
        "verb": "unlabeled",
        "kind": "dropout",
        "count": 58,
        "trailing_median": 445.0,
        "ratio": 0.1303370786516854
      },
      {
        "period": "2021-08",
        "verb": "pr_merged",
        "kind": "dropout",
        "count": 9,
        "trailing_median": 66.0,
        "ratio": 0.13636363636363635
      },
      {
        "period": "2024-02",
        "verb": "ready_for_review",
        "kind": "dropout",
        "count": 4,
        "trailing_median": 23.5,
        "ratio": 0.1702127659574468
      }
    ],
    "volume_flags": [
      {
        "period": "2024-01",
        "n_events": 6288.0,
        "trailing_median": 11912.5,
        "ratio": 0.5278488982161595
      },
      {
        "period": "2024-02",
        "n_events": 5702.0,
        "trailing_median": 11912.5,
        "ratio": 0.47865687303252885
      },
      {
        "period": "2026-08",
        "n_events": 20501.0,
        "trailing_median": 13284.0,
        "ratio": 1.543285155073773
      },
      {
        "period": "2026-09",
        "n_events": 338.0,
        "trailing_median": 14036.0,
        "ratio": 0.02408093473924195
      }
    ],
    "trailing_periods": 6,
    "spike_ratio": 4.0
  },
  "split": null
}

drift.txt

[drift] 839,725 events over 81 periods | structural types excluded: none
[drift] ERA SUSPECT 2026-07: 'issue_closed' spike 748 vs trailing median 95 (7.9x)
[drift] ERA SUSPECT 2023-09: 'issue_closed' spike 139 vs trailing median 20 (7.0x)
[drift] ERA SUSPECT 2026-08: 'pr_closed' spike 745 vs trailing median 125.5 (5.9x)
[drift] ERA SUSPECT 2026-05: 'issue_closed' spike 342 vs trailing median 80.5 (4.2x)
[drift] ERA SUSPECT 2020-09: 'unlabeled' spike 401 vs trailing median 96.5 (4.2x)
[drift] ERA SUSPECT 2026-09: 'renamed' dropout 0 vs trailing median 51.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'issue_closed' dropout 2 vs trailing median 306.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'cross_referenced' dropout 11 vs trailing median 1676.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'review_inline_comment' dropout 10 vs trailing median 747 (0.0x)
[drift] ERA SUSPECT 2026-09: 'review_commented' dropout 9 vs trailing median 624 (0.0x)
[drift] ERA SUSPECT 2026-09: 'pr_commit' dropout 44 vs trailing median 2030 (0.0x)
[drift] ERA SUSPECT 2026-09: 'pr_closed' dropout 3 vs trailing median 136.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'review_approved' dropout 22 vs trailing median 947 (0.0x)
[drift] ERA SUSPECT 2026-09: 'issue_comment' dropout 23 vs trailing median 932 (0.0x)
[drift] ERA SUSPECT 2026-09: 'labeled' dropout 56 vs trailing median 2030 (0.0x)
[drift] ERA SUSPECT 2026-09: 'pr_opened' dropout 12 vs trailing median 433.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'head_force_pushed' dropout 14 vs trailing median 496.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'pr_comment' dropout 77 vs trailing median 2595 (0.0x)
[drift] ERA SUSPECT 2026-09: 'pr_merged' dropout 7 vs trailing median 233 (0.0x)
[drift] ERA SUSPECT 2026-09: 'issue_opened' dropout 4 vs trailing median 132 (0.0x)
[drift] ERA SUSPECT 2026-09: 'review_changes_requested' dropout 1 vs trailing median 33 (0.0x)
[drift] ERA SUSPECT 2026-09: 'ready_for_review' dropout 2 vs trailing median 42.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'unlabeled' dropout 29 vs trailing median 613.5 (0.0x)
[drift] ERA SUSPECT 2026-09: 'review_requested' dropout 10 vs trailing median 199.5 (0.1x)
[drift] ERA SUSPECT 2021-02: 'unlabeled' dropout 58 vs trailing median 445 (0.1x)
[drift] ERA SUSPECT 2021-08: 'pr_merged' dropout 9 vs trailing median 66 (0.1x)
[drift] ERA SUSPECT 2024-02: 'ready_for_review' dropout 4 vs trailing median 23.5 (0.2x)
[drift] VOLUME 2024-01: 6,288 events vs trailing median 11,912 (0.53x)
[drift] VOLUME 2024-02: 5,702 events vs trailing median 11,912 (0.48x)
[drift] VOLUME 2026-08: 20,501 events vs trailing median 13,284 (1.54x)
[drift] VOLUME 2026-09: 338 events vs trailing median 14,036 (0.02x)

knowledge_bound.json

{"knowledge_bound": "2026-09-01T14:14:07.846920+00:00"}

5 files, 733.8 MB

Event spine · 1 file733.8 MB
canonical_events.jsonlb28f40f4b768733.8 MB
Receipt · 3 files9 kB
drift.json65db3aa040996 kB
drift.txtb6a2451c5feb3 kB
knowledge_bound.json0b86d67c698555 B
Phase_report · 1 file1 kB
REPORT.mdd9452c9baa871 kB

Events: 839,725 in 1 stream
Span: 1 Jan 2020 → 1 Sept 2026
Source: GitHub events API
Published: 13 Oct 2026

2 modelstrained or exported over it

pooled8_v1 · 13 Oct 2026 node_serving_v1 · 13 Oct 2026

nodejs/node nodejs__node @ asof-2026-08-24

node_dual's latest zero-shot export target; recorded where it is (the archive drive, _vei_out/envs/node-asof-2026-08-24/nodejs__node/events); see REPORT.md

llm_report · REPORT.md

nodejs__node@asof-2026-08-24 — data stage report

Recorded 2026-10-13 as part of the OSS line's migration to the factory. The data already existed and nothing was rematerialized.

Inputs, by identity. GitHub's event API for nodejs/node, materialized into the OSS canonical spine. Node_dual's latest zero-shot export target. The raw captures stay external, on the archive drive and Modal.

Output, by identity. canonical_events.jsonl, 834,974 events from 2020-01-01T15:11:11Z to 2026-08-24T13:59:31Z. The spine is recorded where it already is (the archive drive, _vei_out/envs/node-asof-2026-08-24/nodejs__node/events) and not copied into the registry.

Drift. Ingest-time check (drift.json, drift.txt): 834,974 events over 80 periods | structural types excluded: none; 7 verb flags, 3 volume flags. The OSS runs split each stream at its own time quantile, so no single calendar split applies across the pooled streams, and no split-time check was run here. Its absence is a caveat on any result.

Receipts (3)

drift.json

{
  "era": {
    "n_periods": 80,
    "n_events": 834974,
    "structural_types": [],
    "verb_flags": [
      {
        "period": "2026-07",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 748,
        "trailing_median": 95.0,
        "ratio": 7.873684210526315
      },
      {
        "period": "2023-09",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 139,
        "trailing_median": 20.0,
        "ratio": 6.95
      },
      {
        "period": "2026-05",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 344,
        "trailing_median": 80.5,
        "ratio": 4.273291925465839
      },
      {
        "period": "2020-09",
        "verb": "unlabeled",
        "kind": "spike",
        "count": 401,
        "trailing_median": 96.5,
        "ratio": 4.155440414507772
      },
      {
        "period": "2021-02",
        "verb": "unlabeled",
        "kind": "dropout",
        "count": 58,
        "trailing_median": 445.0,
        "ratio": 0.1303370786516854
      },
      {
        "period": "2021-08",
        "verb": "pr_merged",
        "kind": "dropout",
        "count": 9,
        "trailing_median": 66.0,
        "ratio": 0.13636363636363635
      },
      {
        "period": "2024-02",
        "verb": "ready_for_review",
        "kind": "dropout",
        "count": 4,
        "trailing_median": 23.5,
        "ratio": 0.1702127659574468
      }
    ],
    "volume_flags": [
      {
        "period": "2024-01",
        "n_events": 6288.0,
        "trailing_median": 11904.0,
        "ratio": 0.5282258064516129
      },
      {
        "period": "2024-02",
        "n_events": 5702.0,
        "trailing_median": 11904.0,
        "ratio": 0.4789986559139785
      },
      {
        "period": "2026-07",
        "n_events": 19284.0,
        "trailing_median": 12646.0,
        "ratio": 1.5249090621540409
      }
    ],
    "trailing_periods": 6,
    "spike_ratio": 4.0
  },
  "split": null
}

drift.txt

[drift] 834,974 events over 80 periods | structural types excluded: none
[drift] ERA SUSPECT 2026-07: 'issue_closed' spike 748 vs trailing median 95 (7.9x)
[drift] ERA SUSPECT 2023-09: 'issue_closed' spike 139 vs trailing median 20 (7.0x)
[drift] ERA SUSPECT 2026-05: 'issue_closed' spike 344 vs trailing median 80.5 (4.3x)
[drift] ERA SUSPECT 2020-09: 'unlabeled' spike 401 vs trailing median 96.5 (4.2x)
[drift] ERA SUSPECT 2021-02: 'unlabeled' dropout 58 vs trailing median 445 (0.1x)
[drift] ERA SUSPECT 2021-08: 'pr_merged' dropout 9 vs trailing median 66 (0.1x)
[drift] ERA SUSPECT 2024-02: 'ready_for_review' dropout 4 vs trailing median 23.5 (0.2x)
[drift] VOLUME 2024-01: 6,288 events vs trailing median 11,904 (0.53x)
[drift] VOLUME 2024-02: 5,702 events vs trailing median 11,904 (0.48x)
[drift] VOLUME 2026-07: 19,284 events vs trailing median 12,646 (1.52x)

knowledge_bound.json

{"knowledge_bound": "2026-08-24T14:06:36.648532+00:00"}

5 files, 728.9 MB

Event spine · 1 file728.9 MB
canonical_events.jsonl172ebb95a38a728.9 MB
Receipt · 3 files3 kB
drift.json126c025d8e042 kB
drift.txt528c72994ba8871 B
knowledge_bound.json5e73440e5f7455 B
Phase_report · 1 file1 kB
REPORT.md2466f0b694751 kB

Events: 834,974 in 1 stream
Span: 1 Jan 2020 → 24 Aug 2026
Source: GitHub events API
Published: 13 Oct 2026

0 modelsnothing trained over it yet

nodejs/node nodejs__node @ train-asof-2026-08-17

the Node serving fine-tune's and node_dual's training stream (829,561 events); recorded where it is (the archive drive, _vei_out/envs/node-asof-2026-08-17/nodejs__node/events); see REPORT.md

llm_report · REPORT.md

nodejs__node@train-asof-2026-08-17 — data stage report

Recorded 2026-10-13 as part of the OSS line's migration to the factory. The data already existed and nothing was rematerialized.

Inputs, by identity. GitHub's event API for nodejs/node, materialized into the OSS canonical spine. The Node serving fine-tune's and node_dual's training stream (829,561 events). The raw captures stay external, on the archive drive and Modal.

Output, by identity. canonical_events.jsonl, 829,561 events from 2020-01-01T15:11:11Z to 2026-08-17T13:31:25Z. The spine is recorded where it already is (the archive drive, _vei_out/envs/node-asof-2026-08-17/nodejs__node/events) and not copied into the registry.

Drift. Ingest-time check (drift.json, drift.txt): 829,561 events over 80 periods | structural types excluded: none; 7 verb flags, 2 volume flags. The OSS runs split each stream at its own time quantile, so no single calendar split applies across the pooled streams, and no split-time check was run here. Its absence is a caveat on any result.

Receipts (3)

drift.json

{
  "era": {
    "n_periods": 80,
    "n_events": 829561,
    "structural_types": [],
    "verb_flags": [
      {
        "period": "2026-07",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 748,
        "trailing_median": 95.0,
        "ratio": 7.873684210526315
      },
      {
        "period": "2023-09",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 139,
        "trailing_median": 20.0,
        "ratio": 6.95
      },
      {
        "period": "2026-05",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 344,
        "trailing_median": 80.5,
        "ratio": 4.273291925465839
      },
      {
        "period": "2020-09",
        "verb": "unlabeled",
        "kind": "spike",
        "count": 401,
        "trailing_median": 96.5,
        "ratio": 4.155440414507772
      },
      {
        "period": "2021-02",
        "verb": "unlabeled",
        "kind": "dropout",
        "count": 58,
        "trailing_median": 445.0,
        "ratio": 0.1303370786516854
      },
      {
        "period": "2021-08",
        "verb": "pr_merged",
        "kind": "dropout",
        "count": 9,
        "trailing_median": 66.0,
        "ratio": 0.13636363636363635
      },
      {
        "period": "2024-02",
        "verb": "ready_for_review",
        "kind": "dropout",
        "count": 4,
        "trailing_median": 23.5,
        "ratio": 0.1702127659574468
      }
    ],
    "volume_flags": [
      {
        "period": "2024-01",
        "n_events": 6286.0,
        "trailing_median": 11902.5,
        "ratio": 0.5281243436252888
      },
      {
        "period": "2024-02",
        "n_events": 5701.0,
        "trailing_median": 11902.5,
        "ratio": 0.4789750052509977
      }
    ],
    "trailing_periods": 6,
    "spike_ratio": 4.0
  },
  "split": null
}

drift.txt

[drift] 829,561 events over 80 periods | structural types excluded: none
[drift] ERA SUSPECT 2026-07: 'issue_closed' spike 748 vs trailing median 95 (7.9x)
[drift] ERA SUSPECT 2023-09: 'issue_closed' spike 139 vs trailing median 20 (7.0x)
[drift] ERA SUSPECT 2026-05: 'issue_closed' spike 344 vs trailing median 80.5 (4.3x)
[drift] ERA SUSPECT 2020-09: 'unlabeled' spike 401 vs trailing median 96.5 (4.2x)
[drift] ERA SUSPECT 2021-02: 'unlabeled' dropout 58 vs trailing median 445 (0.1x)
[drift] ERA SUSPECT 2021-08: 'pr_merged' dropout 9 vs trailing median 66 (0.1x)
[drift] ERA SUSPECT 2024-02: 'ready_for_review' dropout 4 vs trailing median 23.5 (0.2x)
[drift] VOLUME 2024-01: 6,286 events vs trailing median 11,902 (0.53x)
[drift] VOLUME 2024-02: 5,701 events vs trailing median 11,902 (0.48x)

knowledge_bound.json

{"knowledge_bound": "2026-08-17T14:04:20.400418+00:00"}

5 files, 723.6 MB

Event spine · 1 file723.6 MB
canonical_events.jsonl5976166b411c723.6 MB
Receipt · 3 files3 kB
drift.jsonfe9487c427712 kB
drift.txt7bcf498890c7799 B
knowledge_bound.json1d1a63e2ba6255 B
Phase_report · 1 file1 kB
REPORT.md127b9dcfc0c21 kB

Events: 829,561 in 1 stream
Span: 1 Jan 2020 → 17 Aug 2026
Source: GitHub events API
Published: 13 Oct 2026

2 modelstrained or exported over it

node_serving_v1 · 13 Oct 2026 node_dual_v1 · 13 Oct 2026

nodejs/node nodejs__node @ since2020-asof-2026-06-26

pooled8's Node member: the June capture of events created since 2020-01-01 (739,702 events, the count pooled8's receipt implies); recorded where it is (the archive drive, ingest-data/oss_events/nodejs__node); see REPORT.md

llm_report · REPORT.md

nodejs__node@since2020-asof-2026-06-26 — data stage report

Recorded 2026-10-13 as part of the OSS line's migration to the factory. The data already existed and nothing was rematerialized.

Inputs, by identity. GitHub's event API for nodejs/node, materialized into the OSS canonical spine. Pooled8's Node member: the June capture of events created since 2020-01-01 (739,702 events, the count pooled8's receipt implies). The raw captures stay external, on the archive drive and Modal.

Output, by identity. canonical_events_created_since_2020-01-01.jsonl, 739,702 events from 2020-01-01T15:42:17Z to 2026-06-25T19:58:27Z. The spine is recorded where it already is (the archive drive, ingest-data/oss_events/nodejs__node) and not copied into the registry.

Drift. Ingest-time check (drift.json, drift.txt): 739,702 events over 78 periods | structural types excluded: none; 5 verb flags, 2 volume flags. The OSS runs split each stream at its own time quantile, so no single calendar split applies across the pooled streams, and no split-time check was run here. Its absence is a caveat on any result.

Receipts (2)

drift.json

{
  "era": {
    "n_periods": 78,
    "n_events": 739702,
    "structural_types": [],
    "verb_flags": [
      {
        "period": "2026-05",
        "verb": "issue_closed",
        "kind": "spike",
        "count": 344,
        "trailing_median": 81.0,
        "ratio": 4.246913580246914
      },
      {
        "period": "2020-09",
        "verb": "unlabeled",
        "kind": "spike",
        "count": 349,
        "trailing_median": 83.0,
        "ratio": 4.204819277108434
      },
      {
        "period": "2021-02",
        "verb": "unlabeled",
        "kind": "dropout",
        "count": 51,
        "trailing_median": 408.5,
        "ratio": 0.12484700122399021
      },
      {
        "period": "2021-08",
        "verb": "pr_merged",
        "kind": "dropout",
        "count": 9,
        "trailing_median": 66.0,
        "ratio": 0.13636363636363635
      },
      {
        "period": "2024-02",
        "verb": "ready_for_review",
        "kind": "dropout",
        "count": 4,
        "trailing_median": 23.0,
        "ratio": 0.17391304347826086
      }
    ],
    "volume_flags": [
      {
        "period": "2024-01",
        "n_events": 5765.0,
        "trailing_median": 10983.0,
        "ratio": 0.5249021214604389
      },
      {
        "period": "2024-02",
        "n_events": 5356.0,
        "trailing_median": 10983.0,
        "ratio": 0.4876627515250842
      }
    ],
    "trailing_periods": 6,
    "spike_ratio": 4.0
  },
  "split": null
}

drift.txt

[drift] 739,702 events over 78 periods | structural types excluded: none
[drift] ERA SUSPECT 2026-05: 'issue_closed' spike 344 vs trailing median 81 (4.2x)
[drift] ERA SUSPECT 2020-09: 'unlabeled' spike 349 vs trailing median 83 (4.2x)
[drift] ERA SUSPECT 2021-02: 'unlabeled' dropout 51 vs trailing median 408.5 (0.1x)
[drift] ERA SUSPECT 2021-08: 'pr_merged' dropout 9 vs trailing median 66 (0.1x)
[drift] ERA SUSPECT 2024-02: 'ready_for_review' dropout 4 vs trailing median 23 (0.2x)
[drift] VOLUME 2024-01: 5,765 events vs trailing median 10,983 (0.52x)
[drift] VOLUME 2024-02: 5,356 events vs trailing median 10,983 (0.49x)

4 files, 547.2 MB

Event spine · 1 file547.2 MB
canonical_events_created_since_2020-01-01.jsonl5c15cf11c75c547.2 MB
Receipt · 2 files2 kB
drift.json9ad1ce4830a41 kB
drift.txt087b53bc0107629 B
Phase_report · 1 file1 kB
REPORT.mdb7bf1af2ed1c1 kB

Events: 739,702 in 1 stream
Span: 1 Jan 2020 → 25 Jun 2026
Source: GitHub events API
Published: 13 Oct 2026

0 modelsnothing trained over it yet

nodejs/node nodejs__node @ asof-2026-08-25

the stream the served zero-shot export (ft-pooled8 node, as-of 2026-08-17 weights) encoded its latents over; canonical_events_full_asof20260825T135814.jsonl is a byte-identical copy of the spine and is not listed

Receipts (1)

knowledge_bound.json

{"knowledge_bound": "2026-08-25T14:07:00.793636+00:00"}

2 files, 729.1 MB

Event spine · 1 file729.1 MB
canonical_events.jsonled2898bd0882729.1 MB
Receipt · 1 file55 B
knowledge_bound.jsonccd03cdf1d4255 B

Events: 835,408 in 1 stream
Span: 1 Jan 2020 → 25 Aug 2026
Source: GitHub, as of 25 Aug 2026
Published: 7 Oct 2026

3 modelstrained or exported over it

node_serving_v1 · 7 Oct 2026 node_serving_v1 · 7 Oct 2026 node_actor_v1 · 13 Oct 2026

02 · Features

Featurization

An Atlas recipe declares how events become model inputs.

2 feature recipes

events_v1 pipeline/featurize/oss/recipes/events_v1.yaml

Used by: node_actor_v1

1 runtrained over it

refs32_july pipeline/featurize/oss/recipes/refs32_july.yaml

Channels: references
Used by: node_serving_v1, pooled8_v1, node_dual_v1

6 runstrained over it

03 · Training

Recipes and runs

A Praxis recipe fixes the model, the objective and the split.

4 training recipes

node_actor_v1 action-conditioned JEPA · 32-event context · features events_v1 · heads next action 1 run: 13 Oct 2026
node_dual_v1 action-conditioned JEPA · 32-event context · features refs32_july · heads attention 24h, merge 7d, next action · from pooled_rust_node_kubernetes_nixpkgs_airflow_bitcoin_ghost_opencv__action_jepa__drivers_v3__ctx32__modal-pooled8-reffeat32k-anchor__20260729-101817 1 run: 13 Oct 2026
node_serving_v1 action-conditioned JEPA · 32-event context · features refs32_july · heads next action · from pooled_rust_node_kubernetes_nixpkgs_airflow_bitcoin_ghost_opencv__action_jepa__drivers_v3__ctx32__modal-pooled8-reffeat32k-anchor__20260729-101817 4 runs: 13 Oct 2026, 13 Oct 2026, 7 Oct 2026, 7 Oct 2026
pooled8_v1 action-conditioned JEPA · 32-event context · features refs32_july · heads next action 1 run: 13 Oct 2026

7 runs

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof2026081720260817-130359zeroshot_node_asof20260901T135247

21 files, 2.05 GB

Latent matrices · 3 files1.48 GB
action_next_state_latents.npy525.4 MB
action_transition_latents.npy525.4 MB
latents.npy429.9 MB
Auxiliary · 6 files396.6 MB
action_sample_meta.json388.5 MB
aux_next_action_classes.json785 B
aux_next_action_head.pt20 kB
aux_next_action_labels.npy8.1 MB
progress.jsonl21 kB
zeroshot_provenance.json601 B
Sample and event ids · 2 files141.5 MB
action_sample_ids.json93.2 MB
latent_event_ids.json48.3 MB
Checkpoints · 3 files26.9 MB
action_jepa.pt5.4 MB
checkpoint.pt16.2 MB
final.pt5.4 MB
Configuration · 5 files931 kB
action_feature_names.json15 kB
asof_config.json567 kB
praxis_import.json251 B
praxis_paths.json1 kB
praxis.json348 kB
Metrics · 1 file6 kB
metrics.json6 kB
Split · 1 file300 B
split.json300 B

Status: imported
Trained on: nodejs__node @ asof-2026-09-01
Origin: manifest inferred from the directory
Published: 13 Oct 2026

2.05 GB21 artifacts

Ask this model ↓

pooled8_v1 pooled_rust_node_kubernetes_nixpkgs_airflow_bitcoin_ghost_opencv__action_jepa__drivers_v3ctx32modal-pooled8-reffeat32k-anchor20260729-101817zeroshot_node_asof20260901T135247

27 files, 2.04 GB

Latent matrices · 3 files1.48 GB
action_next_state_latents.npy525.4 MB
action_transition_latents.npy525.4 MB
latents.npy429.9 MB
Auxiliary · 13 files389.0 MB
action_sample_meta.json388.8 MB
aux_next_action_head.pt20 kB
progress.jsonl187 kB
remote.json148 B
summoning_probe.apache__airflow.json993 B
summoning_probe.bitcoin__bitcoin.json984 B
summoning_probe.kubernetes__kubernetes.json994 B
summoning_probe.NixOS__nixpkgs.json988 B
summoning_probe.nodejs__node.json1 kB
summoning_probe.opencv__opencv.json714 B
summoning_probe.rust-lang__rust.json978 B
summoning_probe.tryghost__ghost.json979 B
zeroshot_provenance.json656 B
Sample and event ids · 2 files141.5 MB
action_sample_ids.json93.2 MB
latent_event_ids.json48.3 MB
Checkpoints · 3 files26.9 MB
action_jepa.pt5.4 MB
checkpoint.pt16.2 MB
final.pt5.4 MB
Configuration · 5 files932 kB
action_feature_names.json15 kB
asof_config.json567 kB
praxis_import.json306 B
praxis_paths.json1 kB
praxis.json348 kB
Metrics · 1 file5 kB
metrics.json5 kB

Status: imported
Trained on: nodejs__node @ asof-2026-09-01
Origin: manifest inferred from the directory
Published: 13 Oct 2026

2.04 GB27 artifacts

Ask this model ↓

node_actor_v1 node__action_jepa__drivers_v3ctx32modal-node-actorscope-40e__20260828-135349

22 files, 1.31 GB

Latent matrices · 4 files975.8 MB
action_initial_state_latents.npy1.2 MB
action_next_state_latents.npy330.7 MB
action_transition_latents.npy330.7 MB
latents.npy313.2 MB
Auxiliary · 6 files224.4 MB
action_sample_meta.json219.1 MB
aux_next_action_classes.json785 B
aux_next_action_head.pt20 kB
aux_next_action_labels.npy5.2 MB
progress.jsonl72 kB
remote.json147 B
Sample and event ids · 2 files93.6 MB
action_sample_ids.json56.6 MB
latent_event_ids.json36.9 MB
Checkpoints · 3 files12.6 MB
action_jepa.pt2.5 MB
checkpoint.pt7.6 MB
final.pt2.5 MB
Configuration · 5 files283 kB
action_feature_names.json10 kB
asof_config.json56 kB
praxis_import.json207 B
praxis_paths.json308 B
praxis.json216 kB
Metrics · 1 file6 kB
metrics.json6 kB
Split · 1 file300 B
split.json300 B

Status: imported
Trained on: nodejs__node @ asof-2026-08-25
Origin: manifest inferred from the directory
Published: 13 Oct 2026

1.31 GB22 artifacts

Ask this model ↓

node_dual_v1 node__action_jepa__drivers_v3ctx32modal-ft-pooled8-dualaux-cut20260808__20260823-161931

27 files, 1.98 GB

Latent matrices · 4 files1.47 GB
action_initial_state_latents.npy14.5 MB
action_next_state_latents.npy518.0 MB
action_transition_latents.npy518.0 MB
latents.npy424.4 MB
Auxiliary · 11 files337.6 MB
action_sample_meta.json313.2 MB
aux_binary_heads.pt4 kB
aux_binary_labels.npz16.2 MB
aux_next_action_classes.json785 B
aux_next_action_head.pt20 kB
aux_next_action_labels.npy8.1 MB
consequence_targets.json4 kB
progress.jsonl18 kB
remote.json147 B
rollout_skill.json4 kB
STRIPPED.md280 B
Sample and event ids · 2 files137.6 MB
action_sample_ids.json89.9 MB
latent_event_ids.json47.7 MB
Checkpoints · 3 files26.9 MB
action_jepa.pt5.4 MB
checkpoint.pt16.2 MB
final.pt5.4 MB
Configuration · 5 files931 kB
action_feature_names.json15 kB
asof_config.json567 kB
praxis_import.json206 B
praxis_paths.json681 B
praxis.json348 kB
Metrics · 1 file6 kB
metrics.json6 kB
Split · 1 file303 B
split.json303 B

Status: imported
Trained on: nodejs__node @ train-asof-2026-08-17
Origin: manifest inferred from the directory
Published: 13 Oct 2026

1.98 GB27 artifacts

Ask this model ↓

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof20260817__20260817-130359

21 files, 2.01 GB

Latent matrices · 4 files1.48 GB
action_initial_state_latents.npy14.5 MB
action_next_state_latents.npy518.4 MB
action_transition_latents.npy518.4 MB
latents.npy424.7 MB
Auxiliary · 5 files364.1 MB
action_sample_meta.json356.0 MB
aux_next_action_classes.json785 B
aux_next_action_head.pt20 kB
aux_next_action_labels.npy8.1 MB
progress.jsonl21 kB
Sample and event ids · 2 files141.4 MB
action_sample_ids.json92.0 MB
latent_event_ids.json49.4 MB
Checkpoints · 3 files26.9 MB
action_jepa.pt5.4 MB
checkpoint.pt16.2 MB
final.pt5.4 MB
Configuration · 5 files931 kB
action_feature_names.json15 kB
asof_config.json567 kB
praxis_import.json209 B
praxis_paths.json857 B
praxis.json348 kB
Metrics · 1 file6 kB
metrics.json6 kB
Split · 1 file300 B
split.json300 B

Status: imported
Trained on: nodejs__node @ train-asof-2026-08-17
Origin: manifest inferred from the directory
Published: 13 Oct 2026

2.01 GB21 artifacts

Ask this model ↓

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof2026081720260817-130359zeroshot_node_asof20260825T135814_layout2

Re-export of the served zero-shot model under the consolidated layout (#959 one checkpoint, #960 derived latents), 2026-10-07;

More about this run

supersedes the earlier export whose two state matrices were stale links into the parent run. Weights from ft-pooled8-node-asof20260817__20260817-130359 (parent pooled8 modal-pooled8-reffeat32k-anchor__20260729-101817); latents over the as-of-2026-08-25 capture linked here.

19 files, 2.02 GB

Latent matrices · 4 files1.49 GB
action_initial_state_latents.npya7a6147dd20214.6 MB
action_next_state_latents.npybb5ec1a28297522.4 MB
action_transition_latents.npybc1c20155dca522.4 MB
latents.npy3a6bc43ce17f427.7 MB
Auxiliary · 3 files386.3 MB
action_sample_meta.jsonafc982b6b4da386.3 MB
aux_next_action_head.ptaaba0760bc6c20 kB
zeroshot_provenance.json2c0c16adf3ae495 B
Sample and event ids · 2 files142.4 MB
action_sample_ids.jsond72777cdd69b92.7 MB
latent_event_ids.jsondb7d9a3df64249.7 MB
Checkpoints · 1 file5.4 MB
action_jepa.pt0074428ccbfb5.4 MB
Configuration · 4 files931 kB
action_feature_names.json8606cc8b26cd15 kB
asof_config.json313a2afb9146567 kB
praxis_paths.json1cf3905e9991857 B
praxis.json5026897d4154348 kB
Evaluation outputs · 2 files79 kB
gnomon/pr_outcomes_v1/consequence_targets.json080ba0f91b243 kB
gnomon/pr_outcomes_v1/results.json127f7ffe665d76 kB
Serving_report · 1 file44 kB
serving_report/2026-10-13T143515Z.json19c6218dc76e44 kB
Metrics · 1 file5 kB
metrics.jsonaf5072a76b125 kB
Split · 1 file294 B
split.jsonf12053523784294 B

Status: imported
Trained on: nodejs__node @ asof-2026-08-25
Origin: manifest inferred from the directory
Published: 7 Oct 2026

2.02 GB19 artifacts

Ask this model ↓

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof2026081720260817-130359zeroshot_node_asof20260825T135814

The run vei oss serves (CLI state 2026-10-07).

More about this run

Zero-shot export: weights from the node fine-tune on the as-of-2026-08-17 capture (ft-pooled8-node-asof20260817__20260817-130359, parent pooled_rust_node_kubernetes_nixpkgs_airflow_bitcoin_ghost_opencv__action_jepa__drivers_v3__ctx32__modal-pooled8-reffeat32k-anchor__20260729-101817), latents and heads encoded over the as-of-2026-08-25 capture linked here. Training capture and parent not yet registered as rows. SUPERSEDED 2026-10-07 by nodejs__node__action_jepa__drivers_v3__ctx32__ft-pooled8-node-asof20260817__20260817-130359__zeroshot_node_asof20260825T135814_layout2: this export's two state matrices were stale links into the parent run (829,561 rows against the served stream's 835,408); the re-export under the consolidated layout corrects that. Kept as a record; not served.

21 files, 2.94 GB

Latent matrices · 5 files2.42 GB
action_next_state_latents.npy48fb7a4ebebf522.4 MB
action_state_latents.npy8c4b176984e1518.4 MB
action_transition_latents.npy1abb04b1f7b6522.4 MB
latents.npy9316ab6a913b427.7 MB
state_latents.npy1d58cb2b94a8424.7 MB
Auxiliary · 4 files358.8 MB
action_sample_meta.json726c2f7e1cc8358.7 MB
aux_next_action_head.ptaaba0760bc6c20 kB
progress.jsonl41a5285e159121 kB
zeroshot_provenance.json4fcb47c619ea601 B
Sample and event ids · 2 files140.8 MB
action_sample_ids.jsond72777cdd69b92.7 MB
latent_event_ids.json13b586fd1db448.1 MB
Checkpoints · 3 files26.9 MB
action_jepa.pt707cc28a23645.4 MB
checkpoint.ptadf471ca87c616.2 MB
final.ptc5da84d9b9885.4 MB
Configuration · 5 files931 kB
action_feature_names.json8606cc8b26cd15 kB
asof_config.json59130dbcad3d567 kB
praxis_import.json9c968071a583251 B
praxis_paths.json50abf69bccfa1 kB
praxis.json5026897d4154348 kB
Metrics · 1 file6 kB
metrics.jsona487deb3a2a66 kB
Split · 1 file300 B
split.json6c2fac188cb1300 B

Status: imported
Trained on: nodejs__node @ asof-2026-08-25
Origin: manifest inferred from the directory
Published: 7 Oct 2026

2.94 GB21 artifacts

Ask this model ↓

04 · Evaluation

How it compares to other methods

uses the world modelother methodcoin flip

AUC is the chance a method ranks a case that happened above one that didn't: 0.5 is a coin flip, 1.0 is perfect.

The models of record and their headline evaluations. Every result includes the ablations and earlier variants.

Contributors likely to return within 90 days

world model + all0.754

world model alone0.705

facts0.750

0.500.600.700.800.90

Pull requests likely to get a maintainer's response within 3 days

world model + all0.744

world model alone0.703

facts0.752

0.500.600.700.800.90

Pull requests likely to merge

within a week

world model + all0.807

world model alone0.798

facts0.730

clock0.512

0.500.600.700.800.90

Pull requests likely to merge

within two weeks

world model + all0.821

world model alone0.805

facts0.751

clock0.524

0.500.600.700.800.90

Pull requests likely to merge

within four weeks

world model + all0.833

world model alone0.817

facts0.770

clock0.534

0.500.600.700.800.90

Pull requests likely to merge

within three months

world model + all0.853

world model alone0.834

facts0.794

clock0.547

0.500.600.700.800.90

How it is graded: 1 eval config

pr_outcomes_v1 eval config · revision 2 · pipeline/infer/oss/recipes/pr_outcomes_v1.yaml

Targets: merge_within_h 7d, 14d, 28d, 91d
maintainer_response
contributor_churn
Arms: clock, facts, latent, both
Heads: logistic
Split: stream quantile at 0.7
Uncertainty: per episode, paired cluster bootstrap
Drift: at the split

22 cellsover 1 run

node_serving_v1

05 · Serving

What it tells you about nodejs/node

As of 25 Aug 2026.

Report

node_serving_v1
generated 13 Oct 2026

nodejs/node

Where to spend review time

Trust: fair held-out merge-within-28d AUC 0.96; held-out merge-within-365d AUC 0.95; time-to-merge Spearman 0.57 (n_test 103306); not yet backtested as a triage

The 758 pull requests open at the last event, sorted by how likely and how soon each is to merge and whether its recent activity is unusual for it.

Ready to merge 109

predicted chance of merging within 28d in the frontier's top third AND estimated time-to-merge in its fastest third

#65450 crypto: correct CCM decryption FIPS error message
merge in 28d 34%est. ~1 hopen ~4 d
finish the review; close it out
#61215 test_runner: replace native methods with primordials
merge in 28d 2%est. ~2 hopen ~8 mo
finish the review; close it out
#64605 doc: fix fsPromises.watch overflow value
merge in 28d 8%est. ~2 hopen ~37 d
finish the review; close it out
#59038 Fix incorrect comment about caching of certificates when using the `--use-openssl-ca` option.
merge in 28d 1%est. ~5 hopen ~14 mo
finish the review; close it out
#59862 benchmark: refactor http benchmarks
merge in 28d 1%est. ~5 hopen ~12 mo
finish the review; close it out

and 104 more

Off script 26

the last 5 events changed this PR's state in ways the model did not expect, well beyond the PR's own earlier baseline (top 10% of deviations on the open frontier) - something new is going on

#63457 crypto: add null checks for OPENSSL_INIT_new()
merge in 28d 45%est. ~21 hopen ~3 mooff script +1.13
investigate: recent events went off this PR's script
#65054 test: move inline kill-spawn-error script to a fixture
merge in 28d 1%est. ~16 dopen ~20 doff script +1.04
investigate: recent events went off this PR's script
#63154 util: add Temporal support to inspect()
merge in 28d 75%est. ~5 dopen ~4 mooff script +0.95
investigate: recent events went off this PR's script
#63570 squash! lint
merge in 28d 7%est. ~3 dopen ~3 mooff script +0.89
investigate: recent events went off this PR's script
#65112 crypto: prevent Hmac.digest() from returning uninitialized memory
merge in 28d 68%est. ~3 dopen ~18 doff script +0.87
investigate: recent events went off this PR's script

and 21 more

Stalling 208

estimated time-to-merge in the frontier's slowest third (but not meeting all the likely-abandoned conditions)

#64477 repl: add block-based history navigation
merge in 28d 0%est. ~6 dopen ~43 d
engage now to unblock (see review-queue)
#63310 tools: add GitHub Action to ping code owners
merge in 28d 1%est. ~6 dopen ~3 mo
engage now to unblock (see review-queue)
#65311 test: prefer expected failures to skips in WPT status files
merge in 28d 3%est. ~6 dopen ~10 d
engage now to unblock (see review-queue)
#59843 doc: add missing api fs cjs code snippet
merge in 28d 0%est. ~6 dopen ~12 mo
engage now to unblock (see review-queue)
#54407 lib: optimize link header regex for case insensitivity
merge in 28d 0%est. ~6 dopen ~25 mo
engage now to unblock (see review-queue)

and 203 more

Never fit the pattern 29

the model found this PR's state changes hard to predict across its whole history (top 10% of the open frontier) - it never followed this repo's usual patterns

#63827 quic: Fix js lint
merge in 28d 0%est. ~4 dopen ~3 molifetime surprise 1.12
investigate: the model finds this PR's whole history hard to predict
#65461 2026-08-26, Version 24.20.0 'Krypton' (LTS)
merge in 28d 2%est. ~3 dopen ~19 dlifetime surprise 1.08
investigate: the model finds this PR's whole history hard to predict
#64498 [v24.x backport] deps: V8: cherry-pick 1158ae719749
merge in 28d 0%est. ~21 hopen ~2 molifetime surprise 1.03
investigate: the model finds this PR's whole history hard to predict
#64784 test: avoid worker in shared value conveyor test
merge in 28d 2%est. ~3 dopen ~29 dlifetime surprise 0.89
investigate: the model finds this PR's whole history hard to predict
#65161 deps: V8: cherry-pick ba4ef8d72cde
merge in 28d 1%est. ~4 dopen ~16 dlifetime surprise 0.79
investigate: the model finds this PR's whole history hard to predict

and 24 more

Likely abandoned 34

predicted chance of merging within 365d in the frontier's bottom third AND estimated time-to-merge in its slowest third AND age in its oldest third

#50507 refactor:Change the order of branches
merge in 365d 0%est. ~6 dopen ~34 mo
consider closing or pinging the author
#59345 lib: adding `--watch-pattern` to support watching files/dirs with glob
merge in 365d 0%est. ~6 dopen ~13 mo
consider closing or pinging the author
#60633 src: use CP_UTF8 for wide file names on win32
merge in 365d 0%est. ~7 dopen ~10 mo
consider closing or pinging the author
#59566 repl: fix ESLint errors in repl module
merge in 365d 0%est. ~7 dopen ~12 mo
consider closing or pinging the author
#61589 doc: add arm64-musl info into platform list
merge in 365d 0%est. ~7 dopen ~7 mo
consider closing or pinging the author

and 29 more

On track 352

none of the above - on track

and 352 more

Caveats and method

COMPOSITE over THIS run: merged-within-horizon AUC 28d 0.96 / 365d 0.95 (heads fit on pseudo-frontier open-PR cohorts) + time-to-merge Spearman 0.57 + state surprise. est_hours is a LOOSE ranking signal - trust the order of PRs, not the hours (its head is fit on resolved PRs only). Buckets are RELATIVE to this repo's current open frontier (see buckets_criteria); the two abnormal buckets are capped at 10 shown each so they don't crowd out the workload buckets - check last_event_at before treating a flagged PR as active. Not yet backtested as a held-out as-of triage eval (precision@k).

Same as vei oss repo-triage, as of 25 Aug 2026.

Pull requests going off course

Trust: not yet backtested a designed signal, validated by case study only; not backtested

Open pull requests whose latest activity is unlike their own history, and ones that have never followed the project's usual course.

#63457 crypto: add null checks for OPENSSL_INIT_new()
went off its own script

model surprise, event by event · last event 25 Aug 2026

This pull request was marked stale on August 19 after 90 days without activity, then quickly drew approvals from tniessen, jasnell (twice, after a rebase to re-run a failed CI job) and avivkeller by August 23. Four CI runs followed between August 23 and 25, and it was still open at the last event, with a 45% predicted chance of merging in 28 days. The flag reflects a real revival rather than trouble; the open question is why it has not landed.
The events around it (5)
1. 23 Aug 2026 pr comment by nodejs-github-bot
2. 23 Aug 2026 review approved by avivkeller
3. 24 Aug 2026 pr comment by nodejs-github-bot
4. 24 Aug 2026 pr comment by nodejs-github-bot
5. 25 Aug 2026 pr comment by nodejs-github-bot
#65054 test: move inline kill-spawn-error script to a fixture
went off its own script

model surprise, event by event · last event 21 Aug 2026

This looks benign: after avivkeller approved it on August 5, trivikr requested CI on August 21, and the bot posted two CI runs and a coverage report saying all changed lines are covered. That is an ordinary step toward landing. The model's 1% chance of merging in 28 days looks low against these events, so treat that figure with caution.
The events around it (5)
1. 21 Aug 2026 labeled request-ci by trivikr
2. 21 Aug 2026 unlabeled request-ci by github-actions
3. 21 Aug 2026 pr comment by nodejs-github-bot
4. 21 Aug 2026 pr comment by codecov
5. 21 Aug 2026 pr comment by nodejs-github-bot
#64124 fs: throw on existing dir in cpSync with errorOnExist
went off its own script

model surprise, event by event · last event 24 Aug 2026

lpinca approved this pull request on June 25, and since then it has gone through 10 CI runs: five requested by its author between June 27 and July 17, and five more from August 22 to 24. The author also added the "review wanted" label on June 29, and it was still open at the last event. A second reviewer or a look at the CI results may be what it needs.
The events around it (5)
1. 22 Aug 2026 pr comment by nodejs-github-bot
2. 23 Aug 2026 pr comment by nodejs-github-bot
3. 23 Aug 2026 pr comment by nodejs-github-bot
4. 24 Aug 2026 pr comment by nodejs-github-bot
5. 24 Aug 2026 pr comment by nodejs-github-bot
#63827 quic: Fix js lint
never followed the usual pattern

model surprise, event by event · last event 15 Aug 2026

The latest events are about 29 commits and a force push by the author on August 15, 28 of the commits within three seconds, which reads as one bulk update of the branch rather than review activity. It has been quiet since then, and the model gives it almost no chance of merging in 28 days. This is a lead only; a glance at the branch would show whether that update was intended.
The events around it (8)
1. 15 Aug 2026 pr commit by martenrichter
2. 15 Aug 2026 pr commit by martenrichter
3. 15 Aug 2026 pr commit by martenrichter
4. 15 Aug 2026 pr commit by martenrichter
5. 15 Aug 2026 pr commit by martenrichter
6. 15 Aug 2026 pr commit by martenrichter
7. 15 Aug 2026 head force pushed by martenrichter
8. 15 Aug 2026 pr commit by martenrichter
#65461 2026-08-26, Version 24.20.0 'Krypton' (LTS)
never followed the usual pattern

model surprise, event by event · last event 22 Aug 2026

This looks benign: it is the release pull request for version 24.20.0, opened by github-actions, labeled release, gathering commits from many authors, with CI and CITGM runs started on August 22. A pull request like that never follows the usual course, so the flag is expected. Its title names August 26, a day after the data ends, so its 2% chance of merging in 28 days should not be taken at face value.
The events around it (5)
1. 22 Aug 2026 review requested by aduh95
2. 22 Aug 2026 pr commit by nodejs-github-bot
3. 22 Aug 2026 head force pushed by aduh95
4. 22 Aug 2026 pr comment by nodejs-github-bot
5. 22 Aug 2026 pr comment by codecov
#64498 [v24.x backport] deps: V8: cherry-pick 1158ae719749
never followed the usual pattern

model surprise, event by event · last event 31 Jul 2026

This looks benign: it is a backport to the v24.x line that aduh95 suggested closing on July 21, and Renegade334 replied and converted it to a draft on July 31. Nothing has happened since. The open decision is whether to close it or keep it, as that comment thread already raises.
The events around it (5)
1. 14 Jul 2026 labeled v24.x by nodejs-github-bot
2. 14 Jul 2026 labeled v8 engine by nodejs-github-bot
3. 21 Jul 2026 pr comment by aduh95
4. 21 Jul 2026 pr comment by Renegade334
5. 31 Jul 2026 converted to draft by Renegade334

Caveats and method

OK: two lists over per-event state surprise (log of model transition residual / persistence residual). chronically_weird = highest lifetime mean - 'always was strange'. most_anomalous = SELF-NORMALIZED recent deviation: trailing-n mean minus the PR's own earlier mean - 'went off its own script'. Both flag model-unexpected state movement, which is not always outcome-relevant (probe heads read a subspace); designed signals validated by case study only. Gates: >= --min-events history (default 8) and >= --min-age-days of pre-tail history (default 7; a brand-new PR has no script to go off of). Check last_event_at. --action switches to the next-event NLL lens.

Same as vei oss weird-prs, as of 25 Aug 2026.

How long merges are taking

Trust: strong observed: opened and merged times of the PRs merged in the stream

Median hours from opening to merge, by the week of the merge, over the 26 weeks up to the last event.

Latest: 126.6 h; peak 187.8 h.

Who does the reviewing

Trust: strong observed: review events in the stream, bots excluded

Each person's share of the reviews given in the 90 days up to the last event.

5051 reviews (approvals, change requests and comments), bots excluded

Ask your own question

idle node_serving_v1 If the model is cold, the first answer takes a few minutes.

Questions this model answers

Model

nodejs/node

Dataset versions

nodejs/node nodejs__node @ asof-2026-09-01

nodejs__node@asof-2026-09-01 — data stage report

nodejs/node nodejs__node @ asof-2026-08-24

nodejs__node@asof-2026-08-24 — data stage report

nodejs/node nodejs__node @ train-asof-2026-08-17

nodejs__node@train-asof-2026-08-17 — data stage report

nodejs/node nodejs__node @ since2020-asof-2026-06-26

nodejs__node@since2020-asof-2026-06-26 — data stage report

nodejs/node nodejs__node @ asof-2026-08-25

Featurization

events_v1 pipeline/featurize/oss/recipes/events_v1.yaml

refs32_july pipeline/featurize/oss/recipes/refs32_july.yaml

Recipes and runs

node_serving_v1 nodejs__node__action_jepa__drivers_v3__ctx32__ft-pooled8-node-asof20260817__20260817-130359__zeroshot_node_asof20260901T135247

pooled8_v1 pooled_rust_node_kubernetes_nixpkgs_airflow_bitcoin_ghost_opencv__action_jepa__drivers_v3__ctx32__modal-pooled8-reffeat32k-anchor__20260729-101817__zeroshot_node_asof20260901T135247

node_actor_v1 node__action_jepa__drivers_v3__ctx32__modal-node-actorscope-40e__20260828-135349

node_dual_v1 node__action_jepa__drivers_v3__ctx32__modal-ft-pooled8-dualaux-cut20260808__20260823-161931

node_serving_v1 nodejs__node__action_jepa__drivers_v3__ctx32__ft-pooled8-node-asof20260817__20260817-130359

node_serving_v1 nodejs__node__action_jepa__drivers_v3__ctx32__ft-pooled8-node-asof20260817__20260817-130359__zeroshot_node_asof20260825T135814_layout2

node_serving_v1 nodejs__node__action_jepa__drivers_v3__ctx32__ft-pooled8-node-asof20260817__20260817-130359__zeroshot_node_asof20260825T135814

How it compares to other methods

Contributors likely to return within 90 days

Pull requests likely to get a maintainer's response within 3 days

Pull requests likely to merge

Pull requests likely to merge

Pull requests likely to merge

Pull requests likely to merge

pr_outcomes_v1 eval config · revision 2 · pipeline/infer/oss/recipes/pr_outcomes_v1.yaml

What it tells you about nodejs/node

Report

Where to spend review time

Ready to merge 109

Off script 26

Stalling 208

Never fit the pattern 29

Likely abandoned 34

On track 352

Pull requests going off course

How long merges are taking

Who does the reviewing

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof2026081720260817-130359zeroshot_node_asof20260901T135247

pooled8_v1 pooled_rust_node_kubernetes_nixpkgs_airflow_bitcoin_ghost_opencv__action_jepa__drivers_v3ctx32modal-pooled8-reffeat32k-anchor20260729-101817zeroshot_node_asof20260901T135247

node_actor_v1 node__action_jepa__drivers_v3ctx32modal-node-actorscope-40e__20260828-135349

node_dual_v1 node__action_jepa__drivers_v3ctx32modal-ft-pooled8-dualaux-cut20260808__20260823-161931

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof20260817__20260817-130359

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof2026081720260817-130359zeroshot_node_asof20260825T135814_layout2

node_serving_v1 nodejsnodeaction_jepa__drivers_v3ctx32ft-pooled8-node-asof2026081720260817-130359zeroshot_node_asof20260825T135814