Thinking Less to Simulate Better: Intuitive Prompting Improves LLM Agents Simulating Individual Social Media Reactions, Including Unfamiliar Content
3.60T1 sourcearXiv cs.MA
Source record
Published by arXiv cs.MA (T1 source). The original is at https://arxiv.org/abs/2609.30563.
Pipeline notes
The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.
SummaryA study profiled eight Serbian participants and tested four LLMs across five prompt conditions against their reactions to 68 social media posts. Intuitive, immediate responses predicted individual reactions better than analytical prompting, even on topics outside the profile, reducing compressed individual differences from seven times to three times the human baseline.
Why it mattersCounterintuitive result: non-deliberative prompting beats reasoning-style prompting for simulating individual users. Useful technique for anyone building synthetic-user agents, and a challenge to default agent design.
Cited by
No citations on record.