2Declining to review for NeurIPS due to policy on sanctioned institutions (opens in new tab)(pastebin.com)5limoce12h ago0
4Classical Chinese Jailbreak Prompt Optimization via Bio-Inspired Search (opens in new tab)(arxiv.org)2limoce3d ago0
5Cursor's 'Composer 2' model is apparently just Kimi K2.5 with RL fine-tuning (opens in new tab)(old.reddit.com)1limoce6d ago1
7SteptronOss: Lightweight, AI-native training framework for large language models (opens in new tab)(github.com)1limoce23d ago0
9Reduce Complexity of Matmul to O(N^2logN) by Asymptotically Optimal Quantum Algo (opens in new tab)(arxiv.org)4limoce1mo ago1
10Learning from context is harder than we thought (opens in new tab)(hy.tencent.com)245limoce1mo ago136
13LongCat-Flash-Lite 100B A3B Technical Report [pdf] (opens in new tab)(huggingface.co)2limoce1mo ago0
14Jinja.cpp: A single-header C++11 Jinja2 template engine for LLM chat templates (opens in new tab)(github.com)1limoce3mo ago0
15Mini-SGLang: A lightweight yet high-performance inference framework for LLM (opens in new tab)(github.com)2limoce3mo ago0
