openreview.net on Hacker News

Large language models develop novel social biases through adaptive exploration(openreview.net)

194 points by paimapi 1 month ago | 98 comments

Position: LLMs Can't Jump(openreview.net)

290 points by theanonymousone 2 months ago | 192 comments

Transformers are inherently succinct(openreview.net)

123 points by brandonb 4 months ago | 36 comments

Flow Map Learning via Nongradient Vector Flow [pdf](openreview.net)

23 points by E-Reverance 6 months ago | 0 comments

30% drop in O1-preview accuracy when Putnam problems are slightly variated(openreview.net)

529 points by optimalsolver 1 year ago | 490 comments

OpenReview paper page reporting that OpenAI's o1-preview loses about 30% accuracy on slightly varied Putnam math competition problems.

Turing Complete Transformers: Two Transformers Are More Powerful Than One(openreview.net)

176 points by georgehill 2 years ago | 67 comments

Word2Vec received 'strong reject' four times at ICLR2013(openreview.net)

316 points by georgehill 2 years ago | 142 comments

Text Embeddings Reveal (Almost) as Much as Text(openreview.net)

51 points by jxmorris12 3 years ago | 8 comments

Large Language Models Are Human-Level Prompt Engineers(openreview.net)

45 points by cainxinth 3 years ago | 25 comments