What Do VLMs NOTICE? A Mechanistic Interpretability Pipeline for Noise-free Text-Image Corruption and Evaluation
Paper
•
2406.16320
•
Published
•
3
Michal Golovanevsky
mgolov
·
AI & ML interests
None yet
Recent Activity
updated
a dataset
12 days ago
mgolov/Visual-Counterfact
new activity
about 2 months ago
mgolov/Visual-Counterfact:Add task category
upvoted
a
paper
about 2 months ago
Pixels Versus Priors: Controlling Knowledge Priors in Vision-Language
Models through Visual Counterfacts
Organizations
authored
2
papers
5 months ago