AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

Universality and Limitations of Prompt Tuning

Yihan Wang, Jatin Chauhan, Wei Wang, Cho-Jui Hsieh

NeurIPS 2023 · Architectures and efficiency › Pruning and efficient tuning · approach: Parameter-efficient transfer analysis · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase soft prompt length beyond a thresholdprompt tuning, single-layer transformer · constructed memorization datasets S1-S3MSE lossNo clear effect
fine-tune the MLP layer instead of prompt tuningsingle-layer transformer · constructed memorization datasets S1-S3MSE lossImproved
use LoRA with width 2 on the MLP layer instead of prompt tuningsingle-layer transformer · constructed memorization datasets S1-S3training MSE: less than 10^{-10}Improved
train prompt longer until loss saturatesprompt tuning LLaMA 7B · SuperGlue RTEsoft prompt spectral normMixed
train prompt longer until loss saturatesprompt tuning LLaMA 7B · WMT14 En-Frsoft prompt spectral normMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy