ModelRefs / Prompt Tuning — AI Glossary

Prompt Tuning — AI Glossary

A PEFT method training a small set of learnable soft tokens prepended to the input embedding, leaving all model weights frozen.

Overview

Prompt tuning (Lester et al. 2021) appends L trainable vectors to the input sequence. Only these vectors are optimized—no model weight changes. At scale (11B+), prompt tuning matches full fine-tuning on SuperGLUE. Smaller models require more tunable tokens. Enables lightweight task adaptation without the risk of catastrophic forgetting.

Reference details

Topictraining
Last reviewed2026-06-24

Primary source

Continue your research

Use these connected ModelRefs sections to compare alternatives, inspect implementation paths, and review the evidence and governance boundaries relevant to Prompt Tuning — AI Glossary.

Frequently asked questions

What is Prompt Tuning?

A PEFT method training a small set of learnable soft tokens prepended to the input embedding, leaving all model weights frozen.

What concepts are related to Prompt Tuning?

Closely related concepts include prefix tuning, adapter, lora.