Ashwin Ramachandran Ben Thompson It worked on our first try! see our post here: lesswrong.com/posts/NAYyHimM…

i think the model is scared too. 😢

So anyway here's a 7-year-old reddit post with 12 upvotes about higher order moments of statistical distributions that's actually really insightful, even for kurtosis.

commoditized LLMs as a judge are not a panacea for eval automated evals will only ever be as good as the data you configure them on for that reason, we are stoked to introduce Sphynx, a fuzz-tester for surfacing challenging, hallucination-inducing questions. these questions

1/ Models will be extraordinarily good at coding, very soon. Research labs are investing more in coding + reasoning improvements than any other domain for the next model generation. Their efforts will bear fruit.