Zaoral @zaoral@fedibird.com

https://x.com/ArnaudPannatier/status/1799055129829839166
めも

https://x.com/biomedicalhacks/status/1798908487537139863
mamba-2

https://twitter.com/NiJinjie/status/1798182749049852411
評価関連

https://huggingface.co/docs/evaluate/index

https://situational-awareness.ai/from-agi-to-superintelligence/

https://note.com/schroneko/n/n58986a1ed219

https://twitter.com/hillbig/status/1797787321825382872
https://arxiv.org/abs/2405.21060
https://goombalab.github.io/blog/2024/mamba2-part1-model/
mamba2

https://plaza.umin.ac.jp/shoei05/index.php/2024/05/28/2610/

https://gigazine.net/news/20240603-perplexity-pages/

https://wired.jp/article/pocket-sized-ai-models-unlock-new-era-of-computing/

https://arxiv.org/abs/2402.17463
https://x.com/AiXsatoshi/status/1796569553100849251
> LLMは学習された文脈の長さを超えると、性能が低下する。~

https://aws.amazon.com/jp/blogs/news/contribution-mitsui-tender-document-analyzer/

https://dl.acm.org/doi/10.1145/3589335.3651233
トモキくん

https://note.com/npaka/n/n0852d542fd43?sub_rt=share_b

https://gigazine.net/news/20240530-china-ai-gpu-moore-threads/

https://ai-data-base.com/archives/69483

https://twitter.com/shanegJP/status/1795545842469720540
性能

https://note.com/aisatoshi/n/n10a2274c5bc1?sub_rt=share_pb

https://zenn.dev/knowledgesense/articles/2b6aa64f27ea89

https://twitter.com/StartupIdeaAI/status/1795105931744944349
world-model.ai

https://world-model.maitrix.org/