Yuxiang Wei (@YuxiangWei9) / X

Yuxiang Wei

173 posts

Yuxiang Wei

@YuxiangWei9

AI Research @Meta MSL. PhD @siebelschool

Mountain View, CA

yuxiang.cs.illinois.edu

Joined October 2021

Pinned
Yuxiang Wei
@YuxiangWei9
Dec 23, 2025
Software agents can self-improve via self-play RL Introducing Self-play SWE-RL (SSR): training a single LLM agent to self-play between bug-injection and bug-repair, grounded in real-world repositories, no human-labeled issues or tests. 🧵
Yuxiang Wei
@YuxiangWei9
Jul 9
let’s push the horizon longer 🚀
Alexandr Wang
@alexandr_wang
Jul 9
1/ muse spark 1.1 is an industry-competitive agentic and coding model. across many agentic evals it rivals gpt-5.5 and opus-4.8. available now through the new meta model api and in meta ai. 🧵
Yuxiang Wei
@YuxiangWei9
Apr 30
Accepted to ICML 2026! Big thanks to all the collaborators 🎉
Yuxiang Wei
@YuxiangWei9
Dec 23, 2025
Software agents can self-improve via self-play RL Introducing Self-play SWE-RL (SSR): training a single LLM agent to self-play between bug-injection and bug-repair, grounded in real-world repositories, no human-labeled issues or tests. 🧵
Yuxiang Wei
@YuxiangWei9
Apr 16
> Aggressive cheating attempts ...attempt to write their own verifiers to check if their cheating attempts would pass an anti-cheating verifier Super interesting work. Also wild that the agents are relentlessly "creative" at gaming the eval.
Justus Mattern
@MatternJustus
Apr 16
Introducing FrontierSWE, an ultra-long horizon coding benchmark. We test agents on some of the hardest technical tasks like optimizing a video rendering library or training a model to predict the quantum properties of molecules. Despite having 20 hours, they rarely succeed
Yuxiang Wei
@YuxiangWei9
Jan 26
Wow this is smart: per-token feedback, no manual reward design, applicable to partial trajectories in principle, all pain points of outcome RL. so curious to see it being applied at scale or combined with existing RL algos
Siyan Zhao
@siyan_zhao
Jan 22
Introducing 💡On-Policy Self-Distillation💡, a simple method that enables LLM to teach itself with dense per-token feedback on its own on-policy generations—achieving 4-8x more token efficiency vs. GRPO and outperforming both GRPO and SFT/Off-Policy Distillation. Key insight:

New to X?