arXivMaggie Mi, Golzar Atefi, Atsuki Yamaguchi, Felix Gers, Aline Villavicencio, Nafise Sadat MoosaviTue, Jun 2, 2026, 8:59 AM PDT
score 16.4
Language models learn idioms through predictability, not just word composition
Original: Rethinking the Idiomaticity Decomposability Hypothesis: Evidence from Distributional Learning
Source: arxiv.org ↗
Writing ELI5 summary…