[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-llms-can-mimic-human-decisions-but-miss-the-spread":10,"sections":34},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":29,"feedback":33,"feedback_at":22,"cost_usd":33,"total_tokens":33},2880,"llms-can-mimic-human-decisions-but-miss-the-spread","LLMs Can Mimic Human Decisions but Miss the Spread","A study of nine behavioral experiments finds LLMs get the headline result right but mismatch how human responses are distributed.","Large language models can pass a pop quiz on human behavior, but they struggle with the full picture.\n\nResearchers tested nine published behavioral-operations experiments — covering topics like decision biases and behavioral regularities — and graded LLMs on two things: whether they reproduced the correct hypothesis-test outcome, and whether their full response distributions matched human data. On the first measure, LLMs did reasonably well. On the second, they often fell short, even the expensive proprietary models. The main culprit was dispersion: LLMs tend to cluster responses more tightly than humans actually do.\n\nThat gap matters because researchers increasingly use LLMs as cheap substitutes for human subjects in lab studies, field experiments, and surveys. If the distributions are off, any analysis that goes beyond \"did the effect exist\" — effect sizes, variance, tail behavior — could be systematically wrong. Two lightweight fixes helped: chain-of-thought prompting and hyperparameter tuning. Notably, tuning sometimes let smaller open-source models match or beat larger proprietary ones.\n\nThe finding lands at an awkward moment. Synthetic human-behavior data is already circulating in economics and social-science research pipelines. Getting the direction of an effect right while missing the distribution is a bit like a weather model that correctly predicts rain but says it will fall in one spot — useful, until it isn't.","[\"ai\",\"research\",\"large-language-models\",\"behavioral-science\"]","2026-06-30T04:00:00.000Z","2026-06-30T14:29:04.614Z","2026-06-30T14:29:07.558Z","published",null,[],"ai",[24,26,27,28],"research","large-language-models","behavioral-science",[30],{"name":31,"url":32},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2510.03310",0,{"sections":35},[36,40,45,50,55,60,65,70,75,80,85,89,94,99],{"name":37,"slug":24,"count":38,"latest_published_at":39},"AI",2590,"2026-07-16T04:00:00.000Z",{"name":41,"slug":42,"count":43,"latest_published_at":44},"Security","security",294,"2026-07-15T19:59:48.000Z",{"name":46,"slug":47,"count":48,"latest_published_at":49},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",158,"2026-07-16T00:02:48.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",122,"2026-07-14T19:46:26.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Consumer Tech","consumer-tech",93,"2026-07-13T13:20:48.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Software","software",70,"2026-07-13T19:52:25.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",66,"2026-07-10T10:29:37.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Dev Tools","dev-tools",59,"2026-07-07T04:00:00.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":86,"slug":87,"count":83,"latest_published_at":88},"Startups","startups","2026-06-29T20:55:50.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"General","general",29,"2026-07-10T22:28:58.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]