hugofry

Karma: 140

An X-Ray is Worth 15 Features: Sparse Autoencoders for Interpretable Radiology Report Generation

hugofry, Ahmed Abdulaal, NMontanaBrown and a-ijishakin

Oct 7, 2024, 8:53 AM

39 points

1 comment5 min readLW link

(arxiv.org)

Towards Multimodal Interpretability: Learning Sparse Interpretable Features in Vision Transformers

hugofryApr 29, 2024, 8:57 PM

93 points

8 comments11 min readLW link

Robustness of Contrast-Consistent Search to Adversarial Prompting

Nandi, i, Jamie Wright, Seamus_F and hugofry

Nov 1, 2023, 12:46 PM

18 points

1 comment7 min readLW link