Added support for Chain of Thought Prompting (CoT) to improve response accuracy and nuance in AI-driven conversations.
LLM responses now exclude text wrapped in ... tags from being read aloud, allowing the AI to 'think' silently when necessary.
The tag used to mark non-verbal thoughts is configurable via the ThinkTag property in LLMContentProcessor.
Added support for Chain of Thought Prompting (CoT) to improve response accuracy and nuance in AI-driven conversations. LLM responses now exclude text wrapped in... tags from being read aloud, allowing the AI to 'think' silently when necessary.
The tag used to mark non-verbal thoughts is configurable via the
ThinkTagproperty inLLMContentProcessor.