Towards eliciting latent knowledge from LLMs with mechanistic interpretability
Paper
•
2505.14352
•
Published
•
9
Daniel Marczak
danielm1405
AI & ML interests
None yet
Recent Activity
upvoted
a
paper
13 days ago
Towards eliciting latent knowledge from LLMs with mechanistic
interpretability
upvoted
a
paper
4 months ago
Precise Parameter Localization for Textual Generation in Diffusion
Models
authored
a paper
4 months ago
No Task Left Behind: Isotropic Model Merging with Common and
Task-Specific Subspaces
Organizations
None yet
danielm1405's activity
upvoted
a
paper
13 days ago
upvoted
3
papers
4 months ago
Precise Parameter Localization for Textual Generation in Diffusion Models
Paper
•
2502.09935
•
Published
•
12
No Task Left Behind: Isotropic Model Merging with Common and Task-Specific Subspaces
Paper
•
2502.04959
•
Published
•
11
SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders
Paper
•
2501.18052
•
Published
•
8