Skip to content

Sparse autoencoders for vision

active 2025-04-202026-03-24 (UTC)

Partial coverage11,418 / 12,327 hourly files (93%) · 2 absent upstream · 906 failed, retryable2025-03-152026-08-10 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
159
Pushes
98
Pull requests
10
Issues
15
Stars
18
Forks
3

Activity over time

Daily event counts in the loaded window

Line chart, 339 days from 2025-04-20 to 2026-03-24. Pushes: 98 total, peak 5 in a day. Pull requests: 10 total, peak 2 in a day. Issues: 15 total, peak 8 in a day. Comments: 3 total, peak 2 in a day. Stars: 18 total, peak 1 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
samuelstevens1199683
MisteFr1000
Link78081010
Fersoil1000
yhli-ml1000
jakesmbeattie1010

Recent activity

Latest issues, pull requests and releases

  • Pull request#32Link78082026-03-24 09:08
  • Issue#30yhli-ml2025-12-09 06:09
    Any chance SAE can interpret vision models like ResNet, CNN, etc
  • Issue#29samuelstevens2025-11-23 22:31
    Issue running inference step
  • Issue comment#28samuelstevens2025-11-23 19:34
    Issue running the visualisation tools - Documentation outdated
  • Issue#27Fersoil2025-11-23 18:01
    not specifying `val-data.layer` does not throw any error
  • Issue#27samuelstevens2025-11-23 17:57
    not specifying `val-data.layer` does not throw any error
  • Issue#27samuelstevens2025-11-23 17:57
    not specifying `val-data.layer` does not throw any error
  • Issue#26samuelstevens2025-11-23 17:53
    `launch.py` import times are slow
  • Issue#28samuelstevens2025-11-23 17:53
    Issue running the visualisation tools - Documentation outdated
  • Issue comment#27samuelstevens2025-11-23 17:52
    not specifying `val-data.layer` does not throw any error
  • Issue#27samuelstevens2025-11-23 17:50
    not specifying `val-data.layer` does not throw any error
  • Issue#28MisteFr2025-11-23 17:31
    Issue running the visualisation tools - Documentation outdated
  • Issue#26samuelstevens2025-11-17 17:46
    `launch.py` import times are slow
  • Pull request#22samuelstevens2025-10-07 21:34
  • Issue#21samuelstevens2025-10-07 21:31
    Use canonical 'dim' parameter instead of 'axis' in objectives.py
  • Pull request#20samuelstevens2025-10-06 17:09
  • Pull request#18samuelstevens2025-10-06 17:07
  • Pull request#18samuelstevens2025-10-04 19:51
  • Issue#17samuelstevens2025-10-04 17:21
    Bug: BatchLimiter incorrectly counts samples
  • Pull request#16jakesmbeattie2025-09-14 19:35
  • Pull request#14samuelstevens2025-07-31 21:40
  • Pull request#15samuelstevens2025-07-31 21:40
  • Issue comment#7samuelstevens2025-07-31 17:34
    Remove pyvips dependency
  • Issue#8samuelstevens2025-07-31 17:24
    Check all links in documentation.
  • Pull request#13samuelstevens2025-06-26 20:27

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 18 stars here means stars gained during the window, not the repo's star count.