2 papers
cs.CL2026
Quantifying and Mitigating Korean Jamo-Level Typographical Vulnerabilities in Large Language Models
Seojin Lee, Hwanhee Lee
Korean introduces an additional typographical perturbation level not captured by ordinary character-level edit models: because syllable blocks are internally composed of sub-charac…
cs.LG2026
Mitigating Cognitive Inertia in Large Reasoning Models via Latent Spike Steering
Seojin Lee, ByeongJeong Kim, Hwanhee Lee
While Large Reasoning Models (LRMs) have achieved remarkable performance by scaling test-time compute, they frequently suffer from Cognitive Inertia, a failure pattern manifesting…