medium10 pts · first answer countschain-of-thoughtstructured-output
Parsing an answer after reasoning
You want the model to reason before deciding, but your code needs only the final label. Which output design works best?
Discussion 0
Join free to post in the discussion.
Related questions
- Prefilling the responsePrompt Engineer
- Chain-of-thought promptingPrompt Engineer
- JSON that sometimes breaksPrompt Engineer
- Handling invalid model outputPrompt Engineer
- Sampling several answersPrompt Engineer
- Invented values in extractionPrompt Engineer
- Strict JSON hurt accuracyPrompt Engineer
- Forced into the wrong labelPrompt Engineer