Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Jesse Mu Profile picture
Jesse Mu
@jayelmnop
CS PhD student @stanfordnlp @StanfordAILab. Language for machine learning, not machine learning for language. he/him
May 4, 2023 • 4 tweets • 2 min read
New LM eval just dropped—Google has no moat?? Image Problem w/ this article is the cited LM evals are misleading: they don't measure frontier capabilities but a very narrow task distr. Claims that closed LMs have no moat must evaluate OSS models on actual knowledge work, not stuff like "name a restaurant"

semianalysis.com/p/google-we-ha…
Nov 21, 2022 • 8 tweets • 6 min read
Excited to present 3 #NeurIPS2022 papers on a trend I've been very excited about recently: blurring the boundaries between language models and RL agents

(+a bonus 4th paper on active learning!)

🧵(0/7)

PS: I'm on the industry job market! The "They're the same picture" The Office meme. Th 1️⃣ Improving Intrinsic Exploration with Language Abstractions

Using language abstractions to guide exploration in RL, e.g. by self-designing a curriculum of increasingly difficult language goals



Also see @ykilcher review:

(1/7)