 |
ML Newsnew | past | best | rss | submit | about |
|
Stories from October 31, 2025 (UTC)
|
| 1. | Reasoning models reason well, until they don't (arxiv.org) |
| 218 points by optimalsolver 338 days ago | hide | past | pdf | 217 comments | |
| 2. | Investigating How Prompt Politeness Affects LLM Accuracy (arxiv.org) |
| 4 points by awb 338 days ago | hide | past | pdf | 2 comments | |
| 3. | Emu3.5: Native Multimodal Models Are World Learners (arxiv.org) |
| 3 points by progbits 338 days ago | hide | past | pdf | 1 comment | |
| 4. | Defeating the Training-Inference Mismatch via FP16 (arxiv.org) |
| 2 points by matt_d 337 days ago | hide | past | pdf | discuss | |
| 5. | Benchmarking On-Device Machine Learning on Apple Silicon with MLX (arxiv.org) |
| 2 points by PaulHoule 337 days ago | hide | past | pdf | discuss | |
| 6. | Chain-of-Thought Hijacking (arxiv.org) |
| 2 points by belter 337 days ago | hide | past | pdf | discuss | |
| 7. | Collective Communication for 100k+ GPUs (arxiv.org) |
| 1 point by charleshn 338 days ago | hide | past | pdf | discuss | |