Andy Lee
- • master student at SIST, ShanghaiTech University
- • research focus: Interpretability, LLM Safety & Eval
- • open-source contributor
Latest Papers
View All →The Question is the Answer: Weak-to-Strong Benchmarking
In Submission
Identifying Good and Bad Neurons for Task-Level Controllable LLMs
In Submission
GameBench: Evaluating Strategic Reasoning Abilities of LLM Agents
LanGame @ NeurIPS 2024
Delta-Influence: Unlearning Poisons via Influence Functions
ATTRIB @ NeurIPS 2024