model · first seen 6 Oct 2026
Moshi
Open full-duplex speech language model for real-time spoken dialogue.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Steering Follows Geometry, Not Labels: Emotion Directions in a Full-Duplex Speech ModelPapers · 6 Oct 2026 · Pulak Kuli
Latest sources
- Steering Follows Geometry, Not Labels: Emotion Directions in a Full-Duplex Speech ModelPapers · 6 Oct 2026 · Pulak Kuli