- if 30 seconds pass since the last token generation abort the request - stop the llama thread to flush any accumulated context
- if 30 seconds pass since the last token generation abort the request - stop the llama thread to flush any accumulated context