Yunyi Shen/申云逸 🐺
@ShenRaphael
I do random statistics stuff at @MIT EECS, little bit diffusion, little bit LLM. Once UW-Madison and PKU. Take wildlife photos. Opinions are mine.
🚨Dropping Just a Handful of Preferences Can Change Top Large Language Model Rankings. 📍 Oral @ ICML 2025 Workshop on Models of Human Feedback for AI Alignment 🔗New work w/ @ShenRaphael, Dennis Wei, and Tamara Broderick openreview.net/forum?id=b9r1s…
🚨Past work shows: dropping just 0.1% of the data can change the conclusions of important studies. We show: Many approximations can fail to catch this. 📢Check out our new TMLR paper (w/ David Burt, @ShenRaphael , Tin Nguyen, and @ta_broderick ) 👇 openreview.net/forum?id=m6EQ6…
We found this out just today while already in the middle of putting supplementary figures together. Extremely frustrating given the already narrow scope and timelines for responding to reviews.
@NeurIPSConf, why take the option to provide figures in the rebuttals away from the authors during the rebuttal period? Grounding the discussion in hard evidential data (like plots) makes resolving disagreements much easier for both the authors and the reviewers. Left: NeurIPS…
🚀 RL is powering breakthroughs in LLM alignment, reasoning, and agentic apps. Are you ready to dive into the RL x LLM frontier? Join us at @aclmeeting ACL’25 tutorial: Inverse RL Meets LLM Alignment this Sunday at Vienna🇦🇹(Jul 27th, 9am) 📄 Preprint at huggingface.co/papers/2507.13…
Welp, somewhat embarrassingly, yesterday this account was briefly compromised and someone posing as me promoted a crypto project supposedly "tokenizing" manim. Thank you to those of you who quickly flagged it as spam and alerted me to the problem. For future reference, it's a…
Actually I want to know how to cope with this... I feel this constantly.
I use Umaru version of me for my Google scholar profile,😂
🐹Himouto! Umaru-chan. The anime celebrates its 10th anniversary this month. Did you watch it ?
Unfortunately I will have to miss ICML half due to internship half due to US visa situation. But if you want to talk about classic stats in LLM, Schrodinger bridge/diffusion in scientific applications or any other interest stuff please reach out!
Happy to share that our paper on "Active Reward Modeling" has been accepted to ICML 2025! #ICML2025 The part I like the most about the project is its simplicity! Huge thanks to my amazing co-authors @ShenRaphael @HolarisSun More to come! For more detailed 🧵 see 👇
sorry to my advisor my next paper will be delayed by 1 hour and 36 minutes
‘POKEMON: THE FIRST MOVIE’ is now free to watch on YouTube 🔗: youtu.be/25pjnjIVbjU?fe…
Subaru and crew return for even more adventures! Re:ZERO -Starting Life in Another World- Season 4 comes to Crunchyroll next year 🔥
i used to pray for the day nerds would be treated like athletes. the day’s finally here, and it’s fucking glorious
I fabricated and characterized the LPF in this post just to see what it looks like. I made it just by following the pattern with my eyes, but surprisingly, I could see LPF-like behavior!
ok it actually works, uggghhh
Want to download everything from today? Whether you want a new desktop background or an image to send to friends and family, we've got you covered! Head to rubinobservatory.org/gallery/collec… and take your pick🌌 (...or just download everything, no need to choose!)
Rubin is coming!!!!
Introducing...your sneak peek at the cosmos captured by @NSF–@doescience Vera C. Rubin Observatory! Can you guess what regions of sky they are? This is just a peek...join us at 11am US EDT for your full First Look at how Rubin will #CaptureTheCosmos! ls.st/rubin-first-lo…