Sequence architectures and attention · MLD-0152
Causal attention
Attention restricting each position to permitted current or earlier positions.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Prevent next-token prediction from seeing future tokens.
Original caveat
Causal here concerns sequence visibility, not causal inference.
Context and aliases
Level: Intermediate. Stages: Training · Inference · Theory.
Tags: DL · SLM · LLM · NLP.
Aliases: Masked self-attention.
Related concepts
Concepts linking here
Original references
- Attention Is All You Need ↗ · Consulted
- Transformers glossary ↗ · Consulted
- PyTorch neural-network layers ↗ · Reference link
- Original reference URL ↗