arxiv:2602.20021
Gabriele Sarti
gsarti
AI & ML interests
Interpretability for generative language models
Recent Activity
liked a model 4 days ago
neuronpedia/jacobian-lens upvoted a paper 4 days ago
PACE: A Proxy for Agentic Capability Evaluation upvoted a paper 7 days ago
Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads