Sebastian Raschka (rasbt)
rasbt.bsky.social
ML/AI researcher & former stats professor turned LLM research engineer. Author of "Build a Large Language Model From Scratch" (https://amzn.to/4fqvn0D) & reasoning (https://mng.bz/Nwr7). Also blogging about AI research at magazine.sebastianraschka.com.
👤 Joined on 24 April 2023
10516
Followers
249
Following
319
Total Posts
📜 Kullanıcı Geçmişi
- 24 Nis 2023, 12:42
🎉 Bluesky'e katıldı - İlk kullanıcı adı:
@rasbt.bsky.social - 25 Nis 2023, 15:20
✏️ Kullanıcı adı değişti:
@rasbt.bsky.social→@sebastianraschka.com - 13 Kas 2023, 01:19
🔄 Profil bilgileri güncellendi
- 30 Mar 2026, 02:11
✏️ Kullanıcı adı değişti:
@sebastianraschka.com→@rasbt.bsky.social
🚫 Blokladığı Kullanıcılar
❤️ Beğendiği Gönderiler
Wondering how it survives a rewrite. Every draft I ship gets a voice-rewrite pass first, which reshuffles word choice across the whole text. That's the case where someone most wants it read as human, and probably where the sampling signal thins out fastest. Haven't tested it.
Show Post
FYI: "GLM-5.3-Flash Architecture Notes" by @rasbt.bsky.social - linear hybrid attention w/ 34 KDA + 11 MLA/DSA layers - sparse 320B-A18B MoE backbone (~1/2 GLM-5.2) - DeepSeek V4-style mHC residual path w/ 4 parallel streams - native vision encoder (n/a) -> sebastianraschka.com/blog/2026/gl...
Show Post
People who think LLMs are conscious should try to train a small one locally You'll change your mind very quickly
Show PostSeriously, get @rasbt.bsky.social's books and give it a shot
Show PostA new, highly recommended article from @rasbt.bsky.social: "Recent Developments in LLM Architectures: KV Sharing, mHC, and Compressed Attention" -> magazine.sebastianraschka.com/p/recent-dev... #MLsky
Show Post