ARC-AGI-Pub 2024 progress: Jeremy Berman and MIT-Cornell team set new public leaderboard records
Original title:2024 Progress on ARC-AGI-Pub
AI summary
ARC Prize announced that at the end of the 2024 ARC Prize competition, two new approaches set records on the ARC-AGI-Pub public leaderboard: Jeremy Berman achieved 53.6% using Evolutionary Test-time Compute inspired by genetic algorithms with Claude Sonnet 3.5, and a joint MIT and Cornell team achieved 47.5% using test-time training (TTT).
Key Takeaways
- ARC Prize announced that at the end of the 2024 ARC Prize competition,two new approaches set…
- 6% using Evolutionary Test-time Compute inspired by genetic algorithms with Claude Sonnet 3.
- 5,and a joint MIT and Cornell team achieved 47.
- 5% using test-time training (TTT).
Quick Facts
- Published
- Updated
- Source
- ARC Prize:官方博客
- Category
- Papers & Research
- AI score
- 63
原文介绍了两种登上 ARC-AGI 公开榜的方案细节与开源代码,读者可以对比进化式测试时计算与测试时训练的思路差异。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: ARC-AGI-Pub 2024 progress: Jeremy Berman and MIT-Cornell team set new public leaderboard records. December 18, 2024. https://new.nat6.net/en/news/cmtyo2p2t03gerog0vl03n1rp/