Skip to content

Issues: jbloomAus/SAELens

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

31 Open 103 Closed
31 Open 103 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

[Bug Report] b_dec_init_method="geometric_median" incompatible with normalize_activations="expected_average_only_in"
#439 opened Mar 1, 2025 by keltin13
1 task done
1
SAE training on a fine tuned model
#437 opened Feb 26, 2025 by Noam-Diamant
Feature Request: Support custom hf_model in config / loaders
#436 opened Feb 25, 2025 by hijohnnylin
[Proposal] Support early stopping in huggingface LLM wrapper
#429 opened Feb 16, 2025 by chanind
1 task done
1
[Proposal] Optimize decoding in training using torch.nn.functional.embedding_bag
#428 opened Feb 16, 2025 by chanind
1 task done
1
[Proposal] Support Anthropic's JumpReLU improvements
#427 opened Feb 11, 2025 by chanind
[Proposal] Document loading existing SAEs for fine-tuning
#426 opened Feb 11, 2025 by chanind
[Bug Report] tutorials import
#419 opened Feb 5, 2025 by Gregpy
1
[Question] High MSE Loss in Released Gemma SAE Model – Am I Using It Correctly?
#410 opened Jan 28, 2025 by yingjiahao14
1
[Proposal] Automatically/Periodically sync pretrained-saes.yaml
#409 opened Jan 23, 2025 by hijohnnylin
1 task done
@anthonyduong9
[Proposal] Upload Steered Models to Huggingface Hub
#407 opened Jan 18, 2025 by AMindToThink
[Question] fail to reproduce the GPT-2-small SAE with provided hyperparameters
#405 opened Jan 16, 2025 by cwyoon-99
8
[Question] More information about the training process: hyperparameters and token number
#396 opened Dec 3, 2024 by TheMrguiller
6
[Bug Report] sh: trace trap error when trying to train an SAE
#392 opened Nov 30, 2024 by skunnavakkam
1 task done
1
[Bug Report] Unusual L0 Behavior with Gemma-2b
#382 opened Nov 22, 2024 by muyo8692
1 task done
1
[Proposal] Rename l1_coefficient to sparsity_coefficient
#360 opened Nov 4, 2024 by chanind
1 task done
1
1
Training time
#353 opened Oct 30, 2024 by merlinarer
1
[Proposal] Allow excluding special tokens when applying SAEs in HookedSAETransformer
#350 opened Oct 24, 2024 by chanind
1 task done
1
[Proposal] Make the SAE Lens API docs more comprehensive.
#319 opened Oct 4, 2024 by jbloomAus
1
[Bug Report] embedding layer models missing from canonical
#316 opened Oct 3, 2024 by dribnet
1
[Bug Report] Unable to train Mamba SAE
#311 opened Oct 1, 2024 by joelburget
1 task done
2
[Question] How to keep the consistency of ActivationStore
#297 opened Sep 21, 2024 by luckyeric320
[Proposal] Allow users to shuffle ActivationCache dataset, rather than shuffling the pieces of the activation cache
#277 opened Sep 3, 2024 by naterush
1 task done
2
[Bug Report] SAE training tutorial metrics do not match linked run
#276 opened Sep 3, 2024 by naterush
1 task done
4
[Bug Report] Log sparsity artifact seems wrong
#197 opened Jun 29, 2024 by ckkissane
2
ProTip! Updated in the last three days: updated:>2025-03-01.