[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-study-finds-amr-augmentation-doesnt-help-llms":10,"sections":35},{"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":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},9200,"study-finds-amr-augmentation-doesnt-help-llms","Study Finds AMR Augmentation Doesn't Help LLMs","Careful reproduction finds AMR augmentation adds complexity to LLMs without improving their grasp of relational meaning in text.","A new paper says one of NLP's favorite add-ons does basically nothing for today's language models.\n\nAbstract Meaning Representation, or AMR, is a formal graph that maps out who did what to whom in a sentence, stripped of grammar and syntax. Researchers tried to reproduce earlier studies that found big performance gains when LLMs were fed AMR alongside plain text. When they standardized hyperparameter selection across all comparisons, the advantage disappeared: text-only models matched or beat the AMR-augmented ones on every task tested. To dig into why, the team built a perplexity-based probe to check whether AMR was teaching the models anything about sentence relationships they did not already know. It was not.\n\nThat matters because AMR has been a staple ingredient in NLP papers for over a decade, and plenty of pipelines still bolt it onto LLMs assuming it sharpens relational reasoning. If the earlier gains were really an artifact of uneven hyperparameter tuning, teams building on those results have been optimizing for a benefit that was never there.\n\nCall it one more entry in the pile of structured-input tricks that look great until someone runs the control group properly.","[\"ai-research\",\"llms\",\"nlp\",\"reproducibility\"]","2026-10-01T04:00:00.000Z","2026-10-02T01:32:26.409Z","2026-10-02T01:32:30.965Z","published",null,[],"ai",[26,27,28,29],"ai-research","llms","nlp","reproducibility",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.40121",0,{"sections":36},[37,40,44,48,53,58,62,67,72,76,81,86,91,96],{"name":38,"slug":24,"count":39,"latest_published_at":18},"AI",5612,{"name":41,"slug":42,"count":43,"latest_published_at":18},"Security","security",815,{"name":45,"slug":46,"count":47,"latest_published_at":18},"Policy","policy",430,{"name":49,"slug":50,"count":51,"latest_published_at":52},"Deals","deals",298,"2026-09-30T21:00:26.000Z",{"name":54,"slug":55,"count":56,"latest_published_at":57},"Hardware","hardware",196,"2026-09-30T13:00:00.000Z",{"name":59,"slug":60,"count":61,"latest_published_at":18},"Science","science",163,{"name":63,"slug":64,"count":65,"latest_published_at":66},"Consumer Tech","consumer-tech",149,"2026-09-30T22:57:11.000Z",{"name":68,"slug":69,"count":70,"latest_published_at":71},"Dev Tools","dev-tools",93,"2026-10-01T02:30:48.000Z",{"name":73,"slug":74,"count":70,"latest_published_at":75},"Software","software","2026-09-30T21:41:11.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Startups","startups",84,"2026-09-30T20:39:09.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":85},"Gaming","gaming",51,"2026-09-30T16:24:30.000Z",{"name":87,"slug":88,"count":89,"latest_published_at":90},"General","general",50,"2026-09-30T21:37:54.000Z",{"name":92,"slug":93,"count":94,"latest_published_at":95},"Reviews","reviews",31,"2026-09-28T14:31:34.000Z",{"name":97,"slug":98,"count":99,"latest_published_at":100},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]