Controlling Backchannels in Streamable Full-duplex Models
TL;DR
A lightweight backchannel head is introduced that predicts, from a full-duplex model's own hidden states, when a backchannel should begin, and once this probability crosses a tunable threshold, a backchannel is force-decoded.
Abstract
Backchannels, brief acknowledgements like"uh-huh"produced while the other party may still be talking, are central to natural conversation, but full-duplex spoken dialogue models rarely model them explicitly. We introduce a lightweight backchannel head that predicts, from a full-duplex model's own hidden states, when a backchannel should begin. Once this probability crosses a tunable threshold, a backchannel is force-decoded. Attached to both a 7B (PersonaPlex) and a 1B (F-Actor) model, it generalizes across scale. Probing confirms the hidden states anticipate real human timing, and generation evaluation shows more frequent, better-timed backchannels. Human raters judge the resulting backchannels on par with real ones.
