ML Newsnew | past | best | rss | submit about
Stories from October 27, 2025 (UTC)
Go back a day, month, or year. Go forward a day.
1. EntropyLong: Effective Long-Context Training via Predictive Uncertainty (arxiv.org)
15 points by PaulHoule 342 days ago | hide | past | pdf | discuss
2. Learning from Abundant User Dissatisfaction in Real-World Preference Learning (arxiv.org)
3 points by PaulHoule 342 days ago | hide | past | pdf | discuss
3. Collective Communication for 100k+ GPUs (arxiv.org)
2 points by mfiguiere 342 days ago | hide | past | pdf | discuss
4. The Fragility of Benchmark Contamination Detection in Reasoning Models (arxiv.org)
2 points by PaulHoule 342 days ago | hide | past | pdf | discuss
5. Words that make language models perceive (arxiv.org)
2 points by canjobear 342 days ago | hide | past | pdf | discuss
6. Modeling and Solving Operations Research Problems with Tool Augmented LLMs (arxiv.org)
2 points by PaulHoule 342 days ago | hide | past | pdf | discuss
7. Language Models Are Injective and Hence Invertible (arxiv.org)
1 point by porridgeraisin 342 days ago | hide | past | pdf | discuss
8. Reasoning with Sampling: Your Base Model Is Smarter Than You Think (arxiv.org)
1 point by oldfuture 342 days ago | hide | past | pdf | discuss
9. Merge and Conquer: Evolutionarily Optimizing AI for 2048 (arxiv.org)
1 point by xianshou 342 days ago | hide | past | pdf | discuss
10. Stuck in the Matrix: Probing Spatial Reasoning in Large Language Models (arxiv.org)
1 point by xianshou 342 days ago | hide | past | pdf | discuss
About | RSS | RSS (all) | HN arXiv