TY  - RPRT
TI  - Divergence-based Safety Measure for Large Language Models via Rational Inattention
AU  - Anh Tung Nguyen
AU  - Quanyan Zhu
PY  - 2026
UR  - https://arxiv.org/abs/2606.29081
ID  - 2606.29081
ER  -