Every commit scored for change-risk against this repo's own history, so 'elevated' means elevated here rather than on some global curve.
Needs review
111 commits sit in this repo's top risk tercile, which is 33% of the 340scored. The cut is drawn against this codebase's own history rather than a global curve, so a quiet repo still fills its top band, and here it starts at 4.4 out of 10. What pushes a commit up is size and spread together: a large change confined to one area scores below a smaller one scattered across a dozen files.
Commit categories over time, read off the subject line. Fixes carry the accent because that is the series this chart exists to show.
Began docs-led, now leaning fix.
Ranked by change-risk, highest first. Priority is a tercile of this repo's own distribution, so a quiet repo still fills its top band.
| # | Commit | Author | When | Lines | Risk | Top driver |
|---|---|---|---|---|---|---|
| 1 | 4014e62c[fix] restore | jingyaogong | 10mo ago | +31.1K -0 | 100%Elevated | more lines added than baseline |
| 2 | 274483cb250426 | jingyaogong | 1y ago | +31.1K -8 | 100%Elevated | more lines added than baseline |
| 3 | 8be42693MiniMind first open source | gongjy | 2y ago | +41.8K -1 | 100%Elevated | more lines added than baseline |
| 4 | 101d7df2[update] minimind-3 | jingyaogong | 5mo ago | +23.4K -22.2K | 99%Elevated | more lines added than baseline |
| 5 | ee218402update some explain of the code | gongjy | 1y ago | +19.9K -337 | 99%Elevated | more lines added than baseline |
| 6 | e8484874[feat] pause-training | jingyaogong | 10mo ago | +1.2K -1.3K | 98%Elevated | more lines added than baseline |
| 7 | 7d877db7[feat] minimind-2510 | jingyaogong | 10mo ago | +2.6K -1.0K | 98%Elevated | more lines added than baseline |
| 8 | 58e3af03add minimind2 | gongjy | 1y ago | +4.8K -40.2K | 98%Elevated | more lines added than baseline |
| 9 | a62faf34250426 | jingyaogong | 1y ago | +1.1K -19.6K | 97%Elevated | more lines added than baseline |
| 10 | 2dceaf4a添加注释，方便学习者快速理解 | Ben | 1y ago | +337 -319 | 97%Elevated | more lines added than baseline |
| 11 | 36159fb2update app | jingyaogong | 10mo ago | +349 -18 | 97%Elevated | more lines added than baseline |
| 12 | c2e67c25update readme | gongjy | 1y ago | +389 -253 | 97%Elevated | more lines added than baseline |
| 13 | 51dcf51c添加了argparse，方便命令行输入参数 | Yu Chengzhang | 1y ago | +206 -219 | 96%Elevated | more lines added than baseline |
| 14 | 61cb61a4update minimind-v1-moe | gongjy | 1y ago | +351 -458 | 96%Elevated | more lines added than baseline |
| 15 | 55fc9617Create LICENSE | jingyaogong | 2y ago | +201 -0 | 96%Elevated | more lines added than baseline |
| 16 | 5020dc9d[update] readme | jingyaogong | 3mo ago | +132 -133 | 95%Elevated | more lines added than baseline |
| 17 | eb96113c[feat] update readme | jingyaogong | 10mo ago | +154 -57 | 95%Elevated | more lines added than baseline |
| 18 | 11d5cadbupdate rlhf | gongjy | 1y ago | +144 -36 | 95%Elevated | more lines added than baseline |
| 19 | ae767068Create CODE_OF_CONDUCT.md | jingyaogong | 2y ago | +128 -0 | 95%Elevated | more lines added than baseline |
| 20 | 7e9621feupdate minimind-v1 update readme | gongjy | 2y ago | +128 -0 | 94%Elevated | more lines added than baseline |
| 21 | cc81925e[feat] update readme | jingyaogong | 10mo ago | +108 -46 | 94%Elevated | more lines added than baseline |
| 22 | dd0292f9update readme | gongjy | 1y ago | +98 -190 | 94%Elevated | more lines added than baseline |
| 23 | 35bed893[feat] update eval-llm | jingyaogong | 10mo ago | +88 -182 | 93%Elevated | more lines added than baseline |
| 24 | f7ffdf1f[update] shuffle data | jingyaogong | 7mo ago | +63 -59 | 93%Elevated | more lines added than baseline |
| 25 | 9d898576[update] aux loss | jingyaogong | 7mo ago | +80 -51 | 93%Elevated | more lines added than baseline |
| 26 | e4b8789dupdate readme | gongjy | 1y ago | +85 -66 | 93%Elevated | more lines added than baseline |
| 27 | 90e8601cupdate minimind-v1 | gongjy | 2y ago | +116 -100 | 93%Elevated | more lines added than baseline |
| 28 | e4807a52[feat] update readme | jingyaogong | 9mo ago | +80 -53 | 92%Elevated | more lines added than baseline |
| 29 | c090b69c[update] align loss | jingyaogong | 7mo ago | +82 -125 | 92%Elevated | more lines added than baseline |
| 30 | 773e451b[fix] bugs | jingyaogong | 3mo ago | +58 -56 | 91%Elevated | more lines added than baseline |
| 31 | eb875da3update readme | gongjy | 1y ago | +70 -44 | 91%Elevated | more lines added than baseline |
| 32 | a8ae3427Update data preprocessing methods | gongjy | 1y ago | +81 -163 | 90%Elevated | more lines added than baseline |
| 33 | 19523a12update readme | gongjy | 1y ago | +61 -6 | 90%Elevated | more lines added than baseline |
| 34 | 846b37d6update contributor | gongjy | 1y ago | +63 -10 | 90%Elevated | more lines added than baseline |
| 35 | 2ccc0722update minimind-v1 update readme | gongjy | 2y ago | +60 -11 | 90%Elevated | more lines added than baseline |
| 36 | fe2f1199update readme | gongjy | 1y ago | +68 -118 | 90%Elevated | more lines added than baseline |
| 37 | 63615100[fix] rollout bugs | jingyaogong | 3mo ago | +44 -42 | 89%Elevated | more lines added than baseline |
| 38 | 939dc8ff[update] readme | jingyaogong | 4mo ago | +52 -24 | 89%Elevated | more lines added than baseline |
| 39 | bf123b58[feat] add args | jingyaogong | 9mo ago | +57 -57 | 89%Elevated | more lines added than baseline |
| 40 | dd7a7ef7update readme | gongjy | 1y ago | +50 -34 | 88%Elevated | more lines added than baseline |
| 41 | 81d24a4f[feat] add compile | jingyaogong | 7mo ago | +36 -0 | 88%Elevated | more lines added than baseline |
| 42 | 4a7c1c49update rlaif | jingyaogong | 1y ago | +45 -0 | 88%Elevated | more lines added than baseline |
| 43 | 7fcc46b3update seed set | jingyaogong | 1y ago | +48 -6 | 88%Elevated | more lines added than baseline |
| 44 | d2f5ef43update lr | gongjy | 1y ago | +59 -43 | 88%Elevated | more lines added than baseline |
| 45 | 06a66d88添加了wandb | Yu Chengzhang | 1y ago | +52 -7 | 88%Elevated | more lines added than baseline |
| 46 | 07364c3f[update] rename train tokenizer | jingyaogong | 7mo ago | +42 -63 | 87%Elevated | more lines added than baseline |
| 47 | ccc190da[feat] data process | jingyaogong | 6mo ago | +41 -11 | 86%Elevated | more lines added than baseline |
| 48 | 151fdf7e[feat] update yarn | jingyaogong | 8mo ago | +53 -19 | 86%Elevated | more lines added than baseline |
| 49 | 1713c241[fix] model device | jingyaogong | 9mo ago | +35 -56 | 86%Elevated | more lines added than baseline |
| 50 | b969382fupdate readme | gongjy | 1y ago | +41 -37 | 86%Elevated | more lines added than baseline |
Two views of the same model: where the cuts fall, and what commit shape lands you above them.
Every scored commit, binned on the raw 0 to 10 score rather than the percentile. Percentile ranks are uniform by construction, so that axis has no shape to draw. The dashed lines are the tercile cuts behind each row's priority pill.
The 200 most recent commits, on their own recency sample rather than the feed above: that defaults to risk-sorted, so reusing it would plot only the top tercile and call it the spread. Big and scattered is what the model penalises. Click a dot to open it.
Repowise tracks change history across 22 files in 2233admin/minimind. In the last 90 days 15 files were touched, 44 times in total, most often train_grpo.py. Every commit is scored for change risk from its size, spread and the history of the files it touches.
