Kamil Deja

KDeja's activity

• All
• Models
• Datasets
• Spaces
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

upvoted a paper 7 months ago

Towards eliciting latent knowledge from LLMs with mechanistic interpretability

Paper • 2505.14352 • Published May 20 • 9

upvoted 3 papers 11 months ago

Precise Parameter Localization for Textual Generation in Diffusion Models

Paper • 2502.09935 • Published Feb 14 • 12

No Task Left Behind: Isotropic Model Merging with Common and Task-Specific Subspaces

Paper • 2502.04959 • Published Feb 7 • 11

SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders

Paper • 2501.18052 • Published Jan 29 • 8