PPaperPicks

Alejandro Maté

3 papers at tracked venues · 1 at CORE A* · active 20242025

Venues

Frequent coauthors

Jorge García-Carrasco
×3

Papers

  1. Extracting Interpretable Task-Specific Circuits from Large Language Models for Faster Inference
  2. Detecting and Understanding Vulnerabilities in Language Models via Mechanistic Interpretability
  3. How does GPT-2 Predict Acronyms? Extracting and Understanding a Circuit via Mechanistic Interpretability