arXivYuanqing Cai, Ziyi Huang, Minhao Liu, Lixin Duan, Wen Li, Yanru ZhangTue, May 19, 2026, 10:15 AM PDT

score 16.5

LLMs fail to notice hidden clues even in explicit instructions

Original: MixRea: Benchmarking Explicit-Implicit Reasoning in Large Language Models

Writing ELI5 summary…