Skip to content
mlmentorship

Book IV · Chapter 2

Tokens, position, and long context

Understand input representation, positional information, and context extension.

Priority
Core
Difficulty
Intermediate
Useful for
RS, RE, MLE, AS
Interview rounds
ML breadth, LLM systems

Read first: Transformer architecture and attention

Chapter contents

4 entries · read in order
  1. 01
    Tokenization: BPE, WordPiece, and the LLM era
    Concept
  2. 02
    RoPE, ALiBi, and modern positional encodings
    Concept
  3. 03
    Rotary position embeddings (RoPE)
    Concept
  4. 04
    Long-context LLMs: training and serving techniques
    Concept