[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-ai-post-training-boosts-accuracy-shrinks-coverage":10,"sections":40},{"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":30,"tags":31,"sources":35,"feedback":39,"feedback_at":22,"cost_usd":39,"total_tokens":39},9543,"ai-post-training-boosts-accuracy-shrinks-coverage","AI Post-Training Boosts Accuracy, Shrinks Coverage","A new study finds RL fine-tuning sharpens AI agents' first-try accuracy but narrows how many problems they can solve given extra tries.","New research puts a price tag on what reinforcement learning does to language models after post-training: more consistency, fewer ideas.\n\nResearchers studied 14 pairs of base and post-trained large language models across four model families and three agentic benchmarks, 42 cases total. Post-training raised single-attempt accuracy, but it narrowed the range of tasks a model could solve when given many attempts. Base models paired with a lightweight prompting setup often solved more total problems than their post-trained counterparts once allowed a large sampling budget, despite much lower first-try accuracy. The team calls this gap the \"sharpening tax\" and argues post-training pushes tasks toward two extremes: always solved or never solved.\n\nThat split matters because production AI agents rarely get one shot. They take multiple turns, call tools, and retry. If post-training trades away a model's capacity to explore different solution paths, teams optimizing only for benchmark accuracy may be capping how much their systems improve with extra compute at run time. The researchers' proposed fix, a sampling method that adjusts temperature per prompt based on estimated difficulty, recovered some lost coverage in two agentic environments without giving up the accuracy gains.\n\nFine-tuning is not a free upgrade. It reshapes what a model can do, and this study is a tally of what gets reshaped away.","[\"ai\",\"reinforcement-learning\",\"llm-research\",\"ai-agents\"]","2026-10-02T04:00:00.000Z","2026-10-02T23:54:39.157Z","2026-10-02T23:54:45.240Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"publisher-r1","publisher",1,"The headline is missing punctuation (likely a colon) between 'AI Post-Training' and 'the Sharpening Tax,' making it read as a garbled run-on rather than a finished title.","resolved","ai",[30,32,33,34],"reinforcement-learning","llm-research","ai-agents",[36],{"name":37,"url":38},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.01509",0,{"sections":41},[42,45,49,53,58,62,66,71,76,81,86,91,96,101],{"name":43,"slug":30,"count":44,"latest_published_at":18},"AI",5895,{"name":46,"slug":47,"count":48,"latest_published_at":18},"Security","security",837,{"name":50,"slug":51,"count":52,"latest_published_at":18},"Policy","policy",438,{"name":54,"slug":55,"count":56,"latest_published_at":57},"Deals","deals",317,"2026-10-01T22:00:00.000Z",{"name":59,"slug":60,"count":61,"latest_published_at":18},"Hardware","hardware",199,{"name":63,"slug":64,"count":65,"latest_published_at":18},"Science","science",171,{"name":67,"slug":68,"count":69,"latest_published_at":70},"Consumer Tech","consumer-tech",155,"2026-10-01T19:54:10.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Dev Tools","dev-tools",96,"2026-10-01T16:57:03.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Software","software",93,"2026-09-30T21:41:11.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Startups","startups",90,"2026-10-01T21:55:22.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"Gaming","gaming",53,"2026-10-02T02:50:39.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":102,"slug":103,"count":104,"latest_published_at":105},"How-To","how-to",7,"2026-10-01T09:00:00.000Z"]