February 15, 2025 front

ML Newsnew | past | best | rss | submit

Stories from February 15, 2025 (UTC)

Go back a day, month, or year. Go forward a day.

1.	Can We Trust AI Benchmarks? A Review of Current Issues in AI Evaluation (arxiv.org)
	23 points by rntn on Feb 15, 2025 \| hide \| past \| pdf \| 4 comments
2.	The hierarchy in HNSW is not necessary in high dimensions (arxiv.org)
	5 points by blaise-muhirwa on Feb 15, 2025 \| hide \| past \| pdf \| 1 comment
3.	Direct Ascent Synthesis: Hidden Generative Capabilities in Discriminative Models (arxiv.org)
	2 points by kynez on Feb 15, 2025 \| hide \| past \| pdf \| discuss
4.	What makes math problems hard for reinforcement learning: a case study (arxiv.org)
	2 points by rntn on Feb 15, 2025 \| hide \| past \| pdf \| discuss
5.	The Danger of Overthinking: The Reasoning-Action Dilemma in Agentic Tasks (arxiv.org)
	2 points by stared on Feb 15, 2025 \| hide \| past \| pdf \| discuss
6.	Geometry of Prompting: Distinct Mechanisms of Task Adaptation in Language Models (arxiv.org)
	2 points by PlatinumSoup on Feb 15, 2025 \| hide \| past \| pdf \| discuss
7.	Competitive Programming with Large Reasoning Models (arxiv.org)
	1 point by xdavidliu on Feb 15, 2025 \| hide \| past \| pdf \| discuss