Micah Carroll

4

Papers

774

Total Citations

1

Affiliations

Affiliations

UC Berkeley

Papers (4)

Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback

On Targeted Manipulation and Deception when Optimizing LLMs for User Feedback

AI Alignment with Changing and Influenceable Reward Functions

Uni[MASK]: Unified Inference in Sequential Decision Problems