Learning to Reason in LLMs by Expectation Maximization
Large language models (LLMs) solve reasoning problems by first generating a rationale and then answering. We formalize reasoning as a latent variable model and derive a …
Nearly Optimal Latent State Decoding in Block MDPs
First theoretical analysis of model estimation and reward-free RL of block MDP, without resorting to function approximation frameworks. Lower bounds and algorithms with …
yassir-jedra