Houjun Liu
jemoka.com
NLP & POMDPs; CS@Stanford; gradient descent enthusiast www: jemoka.com ac: nlp.stanford.edu/~houjun/
👤 Joined on 05 March 2024
299
Followers
611
Following
70
Total Posts
📜 Kullanıcı Geçmişi
- 05 Mar 2024, 00:58
🎉 Bluesky'e katıldı - İlk kullanıcı adı:
@jemoka.bsky.social - 13 Kas 2024, 21:01
✏️ Kullanıcı adı değişti:
@jemoka.bsky.social→@jemoka.com
🚫 Blokladığı Kullanıcılar
❤️ Beğendiği Gönderiler
LMAO, openreview down point 9pm UCT when ICLR is supposed to be releasing. Coincidence?
Show PostIntroducing 𝘁𝗵𝗼𝘂𝗴𝗵𝘁𝗯𝘂𝗯𝗯𝗹𝗲𝘀: a *fully unsupervised* LM for input-adaptive parallel latent reasoning ✅ Learn yourself a reasoning model with normal pretraining ✅ Better perplexity compared to fixed thinking tokens No fancy loss, no chain of thought labels 🚀
Show Post
Happy to see this work accepted to #EMNLP2025! 🎉🎉🎉
Show PostNew Paper Day! For EMNLP findings—in LM red-teaming, we show you have to optimize for **both** perplexity and toxicity for high-probability, hard to filter, and natural attacks!
Show Post