The Inadequacy of Offline LLM Evaluations: Personalization and Model Behavior
Angelina Wang critiques standard offline LLM evaluations, arguing for methods that account for personalization and dynamic user interaction in real-world model behavior.
Assistant Professor
Self-contained public lectures, keynote addresses, and research seminars.
Angelina Wang critiques standard offline LLM evaluations, arguing for methods that account for personalization and dynamic user interaction in real-world model behavior.