Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
The Paradox of Outcome Optimization: A Causal Information-Theoretic Bound on Reasoning Shortcuts in LLMs
Zihan Chen, Yiming Zhang, Wenxiang Geng +2
Large Language Models (LLMs) aligned via outcome-based Reinforcement Learning (RL) frequently exhibit a critical failure mode: they achieve high performance on in-distribution benc…
cs.LG2026
Rethinking RL Evaluation: Can Benchmarks Truly Reveal Failures of RL Methods?
Zihan Chen, Yiming Zhang, Hengguang Zhou +3
Current benchmarks are inadequate for evaluating progress in reinforcement learning (RL) for large language models (LLMs).Despite recent benchmark gains reported for RL, we find th…