• Generalizable Prompt Tuning for Audio-Language Models via Semantic Expansion (Findings)

Jaehyuk Jang*, Wonjun Lee*, Kangwook Ko* and Changick Kim  (* These three authors contributed equally)

• FinHarmBench: Financial Jailbreak Benchmark and Unsupervised Safety Fine-Tuning via Refusal Steering Distillation (Industry Track)

Yubin Choi*, Yujin Yang*, S. Kim*, Seokil Ham, Seungju Cho, J. Son, Youngjun Kwak and Changick Kim  (* These three authors contributed equally)