Submitted by Jie Jin - Batched Speech Decisions Without Decoding: Single-Token Supervision Lets a Frozen LLM Hear Beyond the Transcript Adventists.ai 27 2