188 B
188 B
- Fixed spurious LLM inference being triggered when a function call result arrived while the user was actively speaking. The context frame is now suppressed until the user stops speaking.