Feedback
Append message to the conversation as a com.embabel.chat.UserMessage and re-prompt the LLM in the same loop. Lets weak models get one more chance to produce an answer, bounded by the policy's own retry counter.
Append message to the conversation as a com.embabel.chat.UserMessage and re-prompt the LLM in the same loop. Lets weak models get one more chance to produce an answer, bounded by the policy's own retry counter.