2 papers
cs.CL2026
The Right Source Is Not Enough: Understanding History Interference in Speech-LLM Memory
HongJin Song, Weiqiao Shan, Runwu Shi +11
Speech large language models retain audio history to support multi-turn conversations. Yet remembering an event does not ensure that its information is used only when relevant. We…
cs.SD2026
REVE: Efficient Hallucination Correction for Large Audio-Language Models via Reused Encoder States
Hongjin Song, Jiasheng Kuang, Xinyu Yang +4
Large audio-language models may mention acoustic events that are absent from the input. A separate audio event detector can verify these mentions, but doing so requires a second au…