tweetindex
EN

Zeyuan Allen-Zhu, Sc.D. ✓

@ZeyuanAllenZhu · joined 23 Apr 2010

physics of language models @ Meta (FAIR at MSL, not GenAI or TBD) 🎓:Tsinghua Physics — MIT CSAIL — Princeton/IAS 🏅:IOI x 2 — ICPC — USACO — Codejam — math MCM

27 137Followers
575Following
537Posts total
939.4KViews on collected posts

Latest posts

@ZeyuanAllenZhu @syhw You did great work there, I’m sure you’ll continue to thrive if you can find a new environment that supports basic research. (Will you go back to China? I wasn’t sure how to interpret your artwork). 7.1K views · 18 likes · 0 reposts · 1 replies Open on X →
@ZeyuanAllenZhu @syhw Great work over the last few years, and good luck/have fun with what's next! 7.6K views · 27 likes · 0 reposts · 0 replies Open on X →
@ZeyuanAllenZhu @syhw Thank you for all the knowledge you freely shared with the world ❤️ Best of luck on your next endeavor!! 5.8K views · 24 likes · 0 reposts · 0 replies Open on X →
I've decided it's time to resign from FAIR. 🫡 I'm especially grateful for the compute that made much of my research possible: 400 H100/200 GPUs allocated to me by FAIR; over a thousand H100s borrowed from FAIR Europe's CodeGen team led by Gabriel (@syhw ); thousands more https:/
277.4K views · 1.5K likes · 48 reposts · 72 replies Open on X →
Congrats, @Kimi_Moonshot! 『 Kimi’s Four Commandments』have circulated in the Chinese AI community for months. Many people add their own fifth for comic relief, but the original four were the core. Here's an English translation --- since apparently none can be taken for granted. ht
46.3K views · 361 likes · 32 reposts · 7 replies Open on X →
Not mine; and I’m not convinced fixed Mamba2 dominates GDN 🤔 (with 1 open question at the end) In Physics of LM, Part 4.2, I compared Mamba2 vs GDN (my improved GDN2) across 1–8B, 1T tokens, w/ and w/o Canon layers. Across scales, GDN consistently outperforms Mamba2 on https://
41.5K views · 273 likes · 26 reposts · 2 replies Open on X →
soooo... how many papers do we think are invalidated by this? And now think about how many other bugs there must be in any re-implementations of... basically anything. https://t.co/mtK0KihAvF
269.5K views · 1.1K likes · 58 reposts · 38 replies Open on X →
Tutorial II of Physics of LM — 3rd video (Part 4.2) is out. Synthetic playground now meets 8B / 1T reality. 𝐑𝐞𝐬𝐮𝐥𝐭 𝟔: Linear models are 𝐍𝐎𝐓 long-context solution — 𝐏𝐄𝐑𝐈𝐎𝐃. The long-context story is an illusion: retrieval fails at any context length (short or https://t.co/exL8Uxr
Physics of Language Models: Part 4.2, Canon Layers at Scale where Synthetic Pretraining Resonates in Reality
79.4K views · 494 likes · 85 reposts · 11 replies Open on X →
(1/N)🚀Today we launch two tightly connected milestones in the Physics of LM series: a sharpened Part 4.1 (v2.0) and a brand new Part 4.2 — together forming a clear, reproducible, textbook-style reference for principled architecture research. Part 4.1 introduced a synthetic https
Physics of LM: Part 4.2, Canon Layers at Scale where Synthetic Pretraining Resonates in Reality
204.9K views · 1K likes · 160 reposts · 23 replies Open on X →

Against accounts of the same size

5 posts from the last 90 days, next to the 10K–100K follower range. shown widely, but few of those viewers react.

Median views7 595this account924median for 10K–100K
Reach, %27.99%this account3.62%median for 10K–100K
Engagement, %0.41%this account1.52%median for 10K–100K
MetricThis accountMedian for 10K–100KRatio
Median views per post7 5959248.22×
Reach (views ÷ followers)27.99%3.62%7.73×
Engagement rate0.41%1.52%0.27×

Others in this range →   Compare with another account →   How these benchmarks are built →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.

Views per post

204.9K16 Dec
79.4K13 Jan
269.5K25 Feb
41.5K27 Feb
46.3K18 Jul
277.4K21 Aug
5.8K
7.6K
7.1K

Last 9 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.

Engagement rate per post

0.61%16 Dec
0.76%13 Jan
0.44%25 Feb
0.73%27 Feb
0.87%18 Jul
0.61%21 Aug
0.41%
0.36%
0.27%

Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.

What the audience does

Likes60.9%4 826 in total
Reposts5.2%409 in total
Replies1.9%154 in total
Quotes1.0%80 in total
Bookmarks31.0%2 457 in total

Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.

The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.

Similar accounts