AI Paper Atlas

Research map / Large language models / Instruction and fine-tuning

Parameter-Efficient Fine-Tuning Methods for Pretrained Language Models: A Critical Review and Assessment

Lingling Xu, Haoran Xie, S. Joe Qin, Xiaohui Tao, Fu Lee Wang

TPAMI 2026 · Large language models › Instruction and fine-tuning · approach: Efficient LLM fine-tuning systems · 607 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
replace full fine-tuning with sequential adapterRoBERTa-base · GLUEaverage: 84.00 -> 85.39Improved
replace full fine-tuning with sequential adapterRoBERTa-large · GLUEaverage: 86.98 -> 88.58Improved
replace full fine-tuning with prompt-tuning (0.61M params)RoBERTa-base · GLUEaverage: 84.00 -> 74.56Hurt
replace full fine-tuning with prefix-tuningRoBERTa-base · GLUEaverage: 84.00 -> 80.89Hurt
replace full fine-tuning with ProPELT_Adapter (about 1.50% params)RoBERTa-large · GLUEaverage: 86.98 -> 88.70Improved
replace full fine-tuning with LoRARoBERTa-base · GLUEaverage: 84.00 -> 84.33Improved
replace full fine-tuning with AdaLoRARoBERTa-base · GLUEaverage: 84.00 -> 83.61Hurt
replace full fine-tuning with Child-Tuning_DRoBERTa-large · GLUE MNLIaccuracy: 89.42 -> 35.45Hurt
add MAM adapter (parallel adapter plus prefix-tuning) in place of prefix-tuningRoBERTa-base · GLUEaverage: 80.89 -> 84.26Improved
replace full fine-tuning with (IA)3T5-base · WMT16 En-RoBLEU: 0.16 higher with 0.03% trainable parametersImproved
replace full fine-tuning with LoRAT5-base · WMT16 En-RoBLEU: 0.36 higher with 0.39% trainable parametersImproved
replace full fine-tuning with LoRAT5-large · WMT16 En-RoBLEU: 0.01 lower with 0.32% trainable parametersNo clear effect

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