Twitxy Saved

Luc Juggery

@lucjuggery

DevRel Engineer @Exoscale // Kubernetes trainer // CK{A,AD,S} // Learning & Sharing
Nice, FR luc.run Joined June 2008
180 Following 1.5K Followers 4.3K Posts

Latest posts 5

Luc Juggery @lucjuggery
I've been diving deeper into how an #LLM works during #inference. This article attempts to demystify these concepts by exploring how a prompt is transformed into generated tokens, explaining prefill, decode, hidden states, attention, ...

exoscale.com/blog/llm-from-…
www.exoscale.com Inside an LLM: From Prompt to Tokens Learn how LLM inference transforms a prompt into tokens through prefill and decode, hidden states, attention, MLP, KV cache, and transformer layers.