دانشمندا اومدن یه حرکتی زدن که هوش مصنوعی خودش بیفته به جون خودش و نسل بعدیشو قویتر بار بیاره

دانشمندا اومدن یه حرکتی زدن که هوش مصنوعی خودش بیفته به جون خودش و نسل بعدیشو قویتر بار بیاره. قضیه از این قراره که توی پروژه Terminal-RSI، مدل از موفقیتهای قبلیش درس میگیره و هی با تمرین، خودشو ارتقا میده تا یه چرخه خودبهبودی درستوحسابی راه بيفته. کار تا جایی پیش رفته که فقط توی دو دور تمرین، موفقیت یه مدل ۲۷ میلیاردی توی تست ترمینال از ۵۱.۷ درصد رسیده به ۸۶.۵ درصد و تو نسخه سنگینترش هم از صفر به ۱۹.۸ درصد پرواز کرده و رکورد زده. حالا تیم یوچنگ شی دنبال اینه که ببینه این پیشرفت زنجیرهای تا کجا قراره بتازه و دقیقاً کجا ترمز میکشه.
منبع
Terminal-RSI: Recursive Self-Improving Terminal Agents via Recursive Self-Rewrite and Policy-Refine
Converting verifier outcomes into useful policy updates and partial progress into complete task resolution is crucial for terminal agents. We present Terminal-RSI, which alternates Recursive Self-Rewrite (RSR) of verified trajectories with critic-inheriting Recursive Policy-Refine (RPR). We formulate Turn-Boundary Skip-Observation GAE (TB-GAE) to retain token-specific baselines while applying decay across interaction turns. Family Dense Reward (FDR) calibrates semantic progress across task families; Success-Prioritized Reward (SPR) separates strict success from bounded partial credit. A Warm-up Critic Model, recursive critic handoff, and stage-specific learning rates coordinate the two refinement phases. The five-stage sequence from Qwen3.8-27B base model through RSR-1, RPR-1, RSR-2, and RPR-2 yields Terminal-RSI-27B: 19.84% on Terminal-Bench 4.0, 86.52% on Terminal-Bench 2.1, 16.95% on Terminal-Bench 3.0, and 79.00% on Terminal-Bench Hard. On LHTB, process reward rises from 0.21 to 0.37 and strict completion from 0/46 to 5/46; SWR100 success doubles from 3.00% to 6.00%. By recursively coupling verified trajectory rewriting with critic-inheriting policy refinement, each retained stage consolidates discovered behaviors and progressively advances the agent from partial local repairs to strict terminal verification.
خواندن در سایت منبع

نظرها
هنوز نظری نیست.
بخش نظرها فعلاً در دسترس نیست.