Skip to content
GitLab
Explore
Sign in
Overview
Active
Stale
All
This project is mirrored from
https://github.com/neelnanda-io/TransformerLens
. Pull mirroring updated
Sep 19, 2024
.
demo-colab-compatibility
052b0827
·
tested more notebooks
·
Sep 16, 2024
gh-pages
34d4b20b
·
Deploying to gh-pages from @...
·
Sep 13, 2024
main
default
e64888d9
·
Merge pull request #723 from TransformerLensOrg/dev
·
Sep 13, 2024
dev
87edf1d1
·
Redo of #713 (#722)
·
Sep 12, 2024
revert-713-new_gqa
8f8273dd
·
Revert "Ungrouping GQA"
·
Sep 12, 2024
dependencies-beartype
5a5411d8
·
Merge branch 'dev' into dependencies-beartype
·
Sep 10, 2024
memory-efficient-loading
9871e6d7
·
removed manuall gc collection
·
Aug 21, 2024
refactor-device-selection
25bbe11b
·
Merge branch 'dev' into refactor-device-selection
·
Aug 18, 2024
experiment-memory-management
a61aac58
·
Updated state loading to copy by reference
·
Aug 16, 2024
unified-conversions
e4fc78d2
·
started working on generalized weight conversions
·
Jul 26, 2024
test-arena-content
4138ed69
·
locked torch to 2.1
·
Jul 22, 2024
ArthurConmy-patch-4
dfd843c2
·
Fix typo in `embed.py` docs
·
Jul 13, 2024
ArthurConmy-patch-3
18fe2b67
·
Move the HookedSAE / HookedSAETransformer warning to a less prominent part of the README
·
Jul 12, 2024
experiment-gemma-weight-tying
458111dd
·
added block for tying weights if configured that way to gemma
·
Jul 08, 2024
model-baichuan
2f44d5be
·
tried setting pos weight
·
Jul 06, 2024
gemma-2
bf595075
·
moved check
·
Jul 06, 2024
mlp-cleanup
740eef12
·
added mlp factory to various parts of the code
·
Jul 05, 2024
patch-mixtral-nonsense
3ac05a74
·
reverted MOE
·
Jun 27, 2024
mixtral-test
4a7c7dbb
·
removed bias
·
Jun 21, 2024
mixtral-playing
8b3929f1
·
Merge branch 'dev' into mixtral-playing
·
Jun 21, 2024