June 5, 2025 front

ML Newsnew | past | best | rss | submit

Stories from June 5, 2025 (UTC)

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

1.	From tokens to thoughts: How LLMs and humans trade compression for meaning (arxiv.org)
	124 points by ggirelli on Jun 5, 2025 \| hide \| past \| pdf \| 25 comments
2.	Extreme Super-Resolution via Scale Autoregression and Preference Alignment (arxiv.org)
	14 points by Brajeshwar on Jun 5, 2025 \| hide \| past \| pdf \| 2 comments
3.	What LLMss Don't Talk About: Empirical Study of Moderation & Censorship Practice (arxiv.org)
	4 points by superpupervlad on Jun 5, 2025 \| hide \| past \| pdf \| discuss
4.	RL in Name Only? Analyzing the Structural Assumptions in RL Post-Training (arxiv.org)
	2 points by porridgeraisin on Jun 5, 2025 \| hide \| past \| pdf \| discuss
5.	Predicting Empirical AI Research Outcomes with Language Models (arxiv.org)
	2 points by danielmorozoff on Jun 5, 2025 \| hide \| past \| pdf \| discuss
6.	I-Con: A Unifying Framework for Representation Learning (arxiv.org)
	1 point by aerophilic on Jun 5, 2025 \| hide \| past \| pdf \| 1 comment
7.	Questioning Representational Optimism in Deep Learning (arxiv.org)
	1 point by publicdaniel on Jun 5, 2025 \| hide \| past \| pdf \| 3 comments
8.	Selective Adapter Freezing for Memory-Efficient Fine-Tuning of Language Models (arxiv.org)
	1 point by PaulHoule on Jun 5, 2025 \| hide \| past \| pdf \| discuss