Twitxy Saved

Richard Zhuang

Verified account This account is verified. Learn more

@RichardZ412

Making Claude precisely safe @AnthropicAI |Prev. @Stanford @bespokelabsai @UCBerkeley
California, USA richardzhuang0412.github.io Joined September 2021
853 Following 1.3K Followers 452 Posts

Latest posts 5

Richard Zhuang Verified account @RichardZ412
As model unlock more and more capabilities both vertically (core intelligence) and horizontally (by domain), evaluation efficiency became the next bottleneck.

Harbor-Index is the exact demonstration of how one should build their evaluation suite. Great work by the team!
Lin Shi Verified account @LinShi592021 Jul 8
Introducing Harbor-Index, a compact, diverse, and high-quality benchmark built to challenge frontier agents.

We carefully select, audit and fix 82 high-signal tasks out of 6,627 candidates spanning 54 benchmarks.

No agent gets above 30%. (1/5)
View quoted post on X
3 replies 5 reposts 16 likes 2.6K views View on X
Richard Zhuang Verified account @RichardZ412
Another Datacomp masterpiece!
Vishaal Udandarao @vishaal_urao Jul 2
🚀New Paper
arxiv.org/abs/2606.28551

Everyone obsesses over VLM architectures & training recipes. But what about the data?

Presenting the latest work in the DataComp-series: a testbed for VLM data curation with 1,000+ controlled experiments and some surprising lessons 👀

🧵👇
View quoted post on X