AI TECHNOLOGY
Readable notes on LLMs, tokens, KV Cache, inference bottlenecks, memory movement, and AI accelerator architecture.
A precise, example-led walkthrough of tokenization, position information, Query-Key matching, Value mixing, residual connections, and what the KV cache does and does not save during autoregressive decoding.
AI TECHNOLOGY A six-page classroom that starts with a simple question: how does an LLM keep writing one piece at a time? We begin with next-token prediction, then move into tokens, vectors, K/V notes, KV Cache, and the memory cost that appears when conversations get long.
