Do 'negative prompts' ('do not X') work in LLMs?
easyAnswer
- Sometimes, sometimes counterproductive.
- Instruction-tuned LLMs generally follow 'do not' directives, but there's a well-documented 'ironic rebound' — mentioning the forbidden thing primes it.
- Better patterns: (1) positive rephrase ('respond only in JSON') vs negative ('do not respond in prose'); (2) explicit output schema; (3) format constraints via examples.
- Use negative prompts sparingly and pair with positive examples.
Check yourself — multiple choice
- Negatives always work
- Work sometimes; 'ironic rebound' primes forbidden content; prefer positive rephrase + examples + constraints
- Impossible
- Only for images
Negative prompts: modest effect, can prime forbidden content — prefer positive framing.
#prompting
Practise LLMs & GenAI
214 interview questions in this topic.