ML Newsnew | past | best | rss | submit about
Stories from December 9, 2025 (UTC)
Go back a day, month, or year. Go forward a day.
1. The universal weight subspace hypothesis (arxiv.org)
358 points by lukeplato 299 days ago | hide | past | pdf | 132 comments
2. Optical context compression is just (bad) autoencoding (arxiv.org)
21 points by unclefuzzy 299 days ago | hide | past | pdf | 1 comment
3. Is Vibe Coding Safe? Benchmarking Vulnerability of Agent-Generated Code (arxiv.org)
4 points by flail 299 days ago | hide | past | pdf | 1 comment
4. Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models (arxiv.org)
4 points by mooreds 299 days ago | hide | past | pdf | discuss
5. Words That Make Language Models Perceive (arxiv.org)
3 points by tbruckner 299 days ago | hide | past | pdf | discuss
6. How to Guess a Gradient (arxiv.org)
2 points by tbruckner 299 days ago | hide | past | pdf | discuss
7. Self-Improving VLM Judges Without Human Annotations (arxiv.org)
1 point by gmays 299 days ago | hide | past | pdf | discuss
8. Benchmarking Vulnerability of Agent-Generated Code in Real-World Tasks (arxiv.org)
1 point by doppp 299 days ago | hide | past | pdf | discuss
About | RSS | RSS (all) | HN arXiv