EMNLP Findings2023
On Robustness of Finetuned Transformer-based NLP Models
Pavan Kalyan Reddy Neerudu, Subba Reddy Oota, Mounika Marreddy, Venkateswara Rao Kagita, Manish Gupta
Abstract
Transformer-based pretrained models like BERT, GPT-2 and T5 have been finetuned for a large number of natural language processing (NLP) tasks, and have been shown to be very effective. However, while finetuning, what changes across layers in these models with respect to pretrained checkpoints is under-studied. Further, how robust are these models to perturbations in input text? Does the robustness vary depending on the NLP task for which the models have been finetuned? While there exists some work on studying the robustness of BERT finetuned for a few NLP tasks, there is no rigorous study that compares this robustness across encoder only, decoder only and encoderdecoder models.
In this paper, we characterize changes between pretrained and finetuned language model representations across layers using two metrics: CKA and STIR. Further, we study the robustness of three language models (BERT, GPT-2 and T5) with eight different text perturbations on classification tasks from the General Language Understanding Evaluation (GLUE) benchmark, and generation tasks like summarization, free-form generation and question generation. GPT-2 representations are more robust than BERT and T5 across multiple types of input perturbation. Although models exhibit good robustness broadly, dropping nouns, verbs or changing characters are the most impactful. Overall, this study provides valuable insights into perturbation-specific weaknesses of popular Transformer-based models, which should be kept in mind when passing inputs. We make the code and models publicly available 1 .
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
2 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
Cited by top-tier papers2
Builds on6
- Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and EntailmentDi Jin, Zhijing Jin, Joey Tianyi Zhou, Peter SzolovitsAAAI 2020 · 1,333 citations
- BERT-ATTACK: Adversarial Attack Against BERT Using BERTLinyang Li, Ruotian Ma, Qipeng Guo, Xiangyang Xue et al.EMNLP 2020 · 529 citations
- CommonGen: A Constrained Text Generation Challenge for Generative Commonsense ReasoningBill Yuchen Lin, Wangchunshu Zhou, Ming Shen, Pei Zhou et al.EMNLP Findings 2020 · 239 citations
- Measuring Representational Robustness of Neural Networks Through Shared InvariancesVedant Nanda, Till Speicher, Camila Kolling, John P. Dickerson et al.ICML 2022 · 15 citations
- RobustLR: A Diagnostic Benchmark for Evaluating Logical Robustness of Deductive ReasonersSoumya Sanyal, Zeyi Liao, Xiang RenEMNLP 2022 · 6 citations
Related papers
- Evaluating the Robustness of Neural Language Models to Input PerturbationsMilad Moradi, Matthias SamwaldEMNLP 2021 · 64 citations
- How transfer learning impacts linguistic knowledge in deep NLP models?Nadir Durrani, Hassan Sajjad, Fahim DalviACL Findings 2021 · 39 citations
- On the Interplay Between Fine-tuning and Sentence-Level Probing for Linguistic Knowledge in Pre-Trained TransformersMarius Mosbach, Anna Khokhlova, Michael A. Hedderich, Dietrich KlakowEMNLP Findings 2020 · 16 citations
- HyPe: Better Pre-trained Language Model Fine-tuning with Hidden Representation PerturbationHongyi Yuan, Zheng Yuan, Chuanqi Tan, Fei Huang et al.ACL 2023 · 5 citations
- Remarkable Robustness of LLMs: Stages of Inference?Vedang Lad, Jin Hwa Lee, Wes Gurnee, Max TegmarkNeurIPS 2025 · 5 citations
