1 paper · 1 filter
Manuel Israel Cázares, Wenlin Zhang, Haobo Ma
We present an empirical study of structural routing failure in large language models (LLMs) over a formally verified algebraic corpus. The task requires selecting the correct proof…