While using the Research (Experimental) model I occasionally run into issues where it has finished generating but fails to actually report back to the UI that it has stopped.
My sessions end up getting left in this weird in-between state where the UI thinks the AI is still generating and even if I hit the "stop" button it doesn't stop. The only way around it once it gets in this state is to start a new thread.
If the AI gets stuck in this loop the "Stop" button should stop both the AI generation and tell the web UI that the AI has been "stopped" so that way you can resume inside the same thread with all the thread context intact including whatever the AI generated prior before getting in the stuck state so that if you ask it to "continue" where it left off, it can.
