The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.
Enhancing multi-step reasoning abilities of language models through direct q-function optimization.arXiv preprint arXiv:2410.09302
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.LG 2years
2026 2representative citing papers
rePIRL learns token-level process rewards for LLM reasoning via a guided-cost-learning-style IRL objective, and shows these rewards improve reasoning policies on math/coding benchmarks.
citing papers explorer
-
Fast Rates for Offline Contextual Bandits with Forward-KL Regularization under Single-Policy Concentrability
The paper establishes the first tilde O(epsilon^{-1}) upper bounds and matching lower bounds for forward-KL-regularized offline contextual bandits under single-policy concentrability in both tabular and general function approximation settings.
-
rePIRL: Learn PRM with Inverse RL for LLM Reasoning
rePIRL learns token-level process rewards for LLM reasoning via a guided-cost-learning-style IRL objective, and shows these rewards improve reasoning policies on math/coding benchmarks.