Souper-Model: How Simple Arithmetic Unlocks State-of-the-Art LLM Performance
Paper • 2511.13254 • Published • 140
vlad
nitrox
AI & ML interests
None yet
Recent Activity
liked a model 3 minutes ago
catplusplus/Qwen3-Omni-30B-A3B-Thinking-heretic-NVFP4 liked a model about 16 hours ago
baidu/Unlimited-OCR liked a model 3 months ago
LittleNicky55/MiniMax-M2.7-abliterated-Heretic-FP8Organizations
None yet
upvoted a paper 8 months ago
upvoted an article over 1 year ago
Article
Bamba: Inference-Efficient Hybrid Mamba2 Model
- +27
Linsong-C, divykum, tridao, albertgu, rganti, dakshiagrawal, mudhakar, daviswer, mirinflim, aviros, nelsonspbr, tmhoangt, OfirArviv, per, michalshmu, HaochenShen, Minjia, gabegoodhart, Naigang, nickhugs, JRosenkranz, cliuawesome, adhoq26, cyang49, SukritiSharma, anhuong, jaygala, szawad, RTGordon
• • 62