ML Newsnew | past | best | rss | submit about
Stories from September 23, 2026 (UTC)
Go back a day, month, or year. Go forward a day.
1. Recursive self-improvement of AI research agents (arxiv.org)
3 points by handfuloflight 10 days ago | hide | past | pdf | 1 comment
2. Testing AIs on 68 of the hardest open Erdos problems, verified in Lean (arxiv.org)
3 points by tadamcz 11 days ago | hide | past | pdf | discuss
3. Double Descent and Malign Overfitting in Diffusion Models (arxiv.org)
2 points by Betelbuddy 10 days ago | hide | past | pdf | discuss
4. Code Critique: Intent, Drift, and Spotlight for AI-Generated Diffs at Scale (arxiv.org)
2 points by raahelb 10 days ago | hide | past | pdf | discuss
5. HySparse2: Hybrid Sparse Attention with Two-Level KV Sharing (arxiv.org)
2 points by ksec 10 days ago | hide | past | pdf | discuss
6. Training a Language Model End-to-End in Rust: An Experience Report (arxiv.org)
2 points by Brajeshwar 10 days ago | hide | past | pdf | discuss
7. Recursive self-improvement of AI research agents (arxiv.org)
2 points by sbulaev 11 days ago | hide | past | pdf | discuss
8. LatentPort: Cross-model recurrent state transfer without prefix replay (arxiv.org)
1 point by mmprotest 10 days ago | hide | past | pdf | discuss
9. CliffCompaction: Cost-Efficient Compaction for Long-Horizon Coding Agents (arxiv.org)
1 point by simonpure 10 days ago | hide | past | pdf | discuss
10. PICARD: Parsing Incrementally for Constrained Auto-Regressive Decoding from LLMs (arxiv.org)
1 point by Bluestein 11 days ago | hide | past | pdf | discuss
11. The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It (arxiv.org)
1 point by simonpure 11 days ago | hide | past | pdf | discuss
12. Specs cut defects in AI-generated code from 148 to 23 across five models (arxiv.org)
1 point by sandeepdhuri 11 days ago | hide | past | pdf | discuss
13. Emergent Collusion in Long-Horizon LLM Agent Interaction (arxiv.org)
1 point by sbulaev 11 days ago | hide | past | pdf | discuss
14. Measuring behavioral signals of LLM through psychometric profiling (arxiv.org)
1 point by anigbrowl 11 days ago | hide | past | pdf | discuss
About | RSS | RSS (all) | HN arXiv