-
Notifications
You must be signed in to change notification settings - Fork 304
Pull requests: TransformerLensOrg/TransformerLens
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix that if use_past_kv_cache is set to True models from the Bloom family produce weird outputs.
#777
opened Nov 9, 2024 by
degenfabian
Loading…
6 tasks done
Set prepend_bos to false by default for Bloom model family
#775
opened Nov 8, 2024 by
degenfabian
Loading…
7 of 8 tasks
fix the bug that attention_mask and past_kv_cache cannot work together
#772
opened Nov 6, 2024 by
yzhhr
Loading…
7 of 10 tasks
Restore consistency of hook_normalized between LayerNorm and RMSNorm
#770
opened Nov 1, 2024 by
degenfabian
Loading…
6 of 7 tasks
Add skip_verbose_naming in add_hook to give an option for skipping the naming
#635
opened Jun 11, 2024 by
verlocks
Loading…
3 of 7 tasks
NanoGPT Conversation did not handle case when there were no biases in model
#629
opened Jun 7, 2024 by
dashstander
Loading…
2 of 7 tasks
Refactor the utilities file into utilities folder
#628
opened Jun 7, 2024 by
starship006
•
Draft
3 of 10 tasks
Make
FactoredMatrix compatible with tensor-like arguments
#599
opened May 17, 2024 by
JasonGross
•
Draft
2 of 10 tasks
revised demo testing to check all demos
#542
opened Apr 15, 2024 by
bryce13950
•
Draft
1 of 10 tasks
Remove FactoredMatrix.py<->utils.py circular dependency
#524
opened Mar 22, 2024 by
ArthurConmy
•
Draft
Make New feature or request
tokenize_and_concatenate work with more datasets
enhancement
#473
opened Dec 28, 2023 by
ArthurConmy
Loading…
3 tasks
(Draft) Add DLA function to utils
#466
opened Dec 16, 2023 by
VasilGeorgiev39
Loading…
3 of 10 tasks
ProTip!
no:milestone will show everything without a milestone.