-
Notifications
You must be signed in to change notification settings - Fork 0
Pull requests: safety-research/crosscoder_emergent_misalignment
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Implement Delphi integration for CrossCoder interpretability (Issue #703)
#705
opened Jul 26, 2025 by
superkaiba
Collaborator
Loading…
Speed up training code with 4 major performance optimizations
#691
opened Jul 25, 2025 by
superkaiba
Collaborator
Loading…
6 tasks done
Speed up top prompts and token processing analysis steps
#690
opened Jul 25, 2025 by
superkaiba
Collaborator
Loading…
Add comprehensive support for Gemma models
#688
opened Jul 25, 2025 by
superkaiba
Collaborator
Loading…
Implement issue #621: Use safety-tooling package for Anthropic API calls
#624
opened Jul 23, 2025 by
superkaiba
Collaborator
Loading…
Add comprehensive perplexity analysis tool
#604
opened Jul 22, 2025 by
superkaiba
Collaborator
Loading…
Add Matryoshka interpretability analysis standalone script
#571
opened Jul 21, 2025 by
superkaiba
Collaborator
Loading…
Implement MatryoshkaSharedCrossCoder: Combine Matryoshka and Shared Features
#563
opened Jul 21, 2025 by
superkaiba
Collaborator
Loading…
4 of 6 tasks
ProTip!
Adding no:label will show everything without a label.