Decoding and inference behavior · MLD-1019
Repetition penalty
A decoding adjustment discouraging reuse of previously generated tokens or sequences.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Reduce repetitive loops in generated text.
Original caveat
It can suppress necessary repetition and harm accuracy.
Context and aliases
Level: Intermediate. Stages: Inference · Evaluation · Deployment.
Tags: SLM · LLM · NLP.
Aliases: None recorded.
Related concepts
Concepts linking here
Original references
- Hugging Face: generation strategies ↗ · Reference link
- Fast Inference from Transformers via Speculative Decoding ↗ · Consulted
- vLLM documentation ↗ · Consulted
- Original reference URL ↗