cd/entity/cross entropy loss· home› entities› cross entropy loss
grep -l @cross entropy loss /news/*.json | wc -l → 1

cross entropy loss

mentions 1 type Person feed RSS

// recent coverage 1 mentions

22:44
2026-09-14
garden.azl.au
artificial-intelligence

Transformer Basics

A technical reference on transformer architecture details that most tokenizers prepend a BOS token, which acts as a fixed landmark or "attention sink," while a space prefix yields the wrong token. The…

// co-occurs with top 4 entities
BOS token 1 MLP 1 attention heads 1 residual stream 1
// topics top 5 topics
artificial intelligence 1 machine learning 1 large language models 1 neural networks 1 natural language processing 1