28.33% 🔥🔥 on SWE-bench Lite!

x.com/i/article/1813…

You can find me and Sayak Ray Chowdhury presenting our work today at Poster Session 3 (Poster #913) – come say hello! #ICML2024

My First Blog! abinezer.github.io/posts/blog-pos… Would love feedback and comments! Thanks to my friends Anush Kini and Roshni Poddar for reviewing!

(1/4) Typical LLM post-training mechanisms have a hard time learning models that can produce diverse responses. To fix this we introduce 𝐃𝐐𝐎 (𝐃𝐢𝐯𝐞𝐫𝐬𝐢𝐭𝐲 𝐐𝐮𝐚𝐥𝐢𝐭𝐲 𝐎𝐩𝐭𝐢𝐦𝐢𝐳𝐚𝐭𝐢𝐨𝐧), a method for post-training LLMs to generate diverse high-quality

There was a time when you owned the software you bought. I play Minecraft most evenings with my son for an hour. Today, the license server was down, and the game wouldn’t start. We just sat there, staring at the launcher, watching it try to verify our license. A game I paid for.