AI.info

Goran Glavaš

Explore Goran Glavaš on AI.info.

• Compositional Steering of Large Language Models with Steering Tokens
• The Curious Case of Factual (Mis)Alignment between LLMs' Short- and Long-Form Answers