Run and analyze LLM decision theory experiments across Newcomb's Problem, Parfit's Hitchhiker, and Transparent Newcomb's.