Parece simples até o entrevistador perguntar "why?". Abaixo, uma estrutura de resposta que dá conta do follow-up.
A token is a unit produced by a model's tokenizer. Text tokens may be word pieces, punctuation, or whitespace. Token counts vary by language, content, and tokenizer, so measure with the model's tokenizer rather than assuming one word or a fixed character count per token.
A context window limits how much input and generated output a model can handle in one computation, subject to separate model-specific input/output limits. Conversation history, retrieved passages, tool results, and multimodal inputs consume that budget. Reserve room for the answer and check the provider's current counting rules.
Ler é o primeiro passo. No app, essa pergunta vira flashcard com repetição espaçada, e o módulo tem quiz e simulado para testar o que ficou.
Treinar essa pergunta no app