Can We Read the Mind of an Audio LLM? A Verbalizable, Multilingual Middle-Layer Workspace
Published in arXiv preprint, 2026, 2026
Recommended citation: Jiajun Fan, Jingyuan Li, Prashanth Gurunath Shivakumar, Qi Luo, Jia-Hong Huang, M. Maruf, Roger Ren, Yile Gu, Rahul Pandey, Ge Liu, Ivan Bulyko. "Can We Read the Mind of an Audio LLM? A Verbalizable, Multilingual Middle-Layer Workspace." arXiv preprint arXiv:2608.24958, 2026. https://arxiv.org/abs/2608.24958
An audio LLM is a black box in a specific way: we see what it says, never what it worked out on the way there, and chain-of-thought monitoring only helps if the model writes its reasoning down. Reading a base Qwen3-Omni with a logit lens at the audio-token positions, the answer to a spoken question becomes legible — in words — in the middle layers, before a single token is emitted.