{"type":"post","id":"t-tech-lab-announces-f-grpo-and-essa-aaaf91","slug":"t-tech-lab-announces-f-grpo-and-essa-aaaf91","title":"T-Tech lab announces F-GRPO and ESSA to cut the cost of LLM fine-tuning","description":"A paper by Daniil Plyusov and coauthors proposes F-GRPO, a modification to group-relative policy optimization intended to keep rare correct responses from being suppressed during reinforcement learning with verifiable rewards.","retrieval_nugget":"A paper by Daniil Plyusov and coauthors proposes F-GRPO, a modification to group-relative policy optimization intended to keep rare correct responses from being suppressed during reinforcement learning with verifiable rewards. In experiments on Qwen2.5-7B, the method improved reported math pass@256 scores across three training algorithms without increasing group size or computational cost.","published_at":"2026-11-09","updated_at":"2026-11-09","record_date":"2026-11-09","date_kind":"published_at","topics":["models","verification"],"entities":["F-GRPO","Qwen2.5-7B","Daniil Plyusov","Alexey Gorbatovski","GRPO","DAPO","CISPO"],"related_patterns":[],"novelty_label":"new_fact","source_url":"https://arxiv.org/abs/2602.06717","source_urls":["https://arxiv.org/abs/2602.06717"],"schema_version":"newruntime-agent-readable-v0.2","stable_id":"post:t-tech-lab-announces-f-grpo-and-essa-aaaf91","status":"published","visuals":[],"editor":{"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]},"routes":{"html":"https://newruntime.com/posts/t-tech-lab-announces-f-grpo-and-essa-aaaf91/","markdown":"https://newruntime.com/posts/t-tech-lab-announces-f-grpo-and-essa-aaaf91.md","json":"https://newruntime.com/posts/t-tech-lab-announces-f-grpo-and-essa-aaaf91.json"},"markdown_url":"https://newruntime.com/posts/t-tech-lab-announces-f-grpo-and-essa-aaaf91.md","next_reads":[{"type":"topic","path":"/topics/models/","url":"https://newruntime.com/topics/models/","title":"Models","reason":"Explore the models topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/models.md","json_url":"https://newruntime.com/topics/models.json"},{"type":"topic","path":"/topics/verification/","url":"https://newruntime.com/topics/verification/","title":"Verification","reason":"Explore the verification topic hub.","media_type":"text/html","markdown_url":"https://newruntime.com/topics/verification.md","json_url":"https://newruntime.com/topics/verification.json"},{"type":"collection","path":"/posts/","url":"https://newruntime.com/posts/","title":"Field Notes","reason":"Open the latest editorial analysis.","media_type":"text/html","json_url":"https://newruntime.com/posts.json"},{"type":"collection","path":"/sources/","url":"https://newruntime.com/sources/","title":"Sources","reason":"See the original sources this archive cites.","media_type":"text/html"}]}