mlsanityData to 7 Oct 2026

Investigating Forgetting in Pre-Trained Representations Through Continual Learning

arXiv 2305.05968 ↗First posted 10 May 2023cs.CLcited by 19

Not on the map. Why?

Abstract (arXiv)

Representation forgetting refers to the drift of contextualized representations during continual training. Intuitively, the representation forgetting can influence the general knowledge stored in pre-trained language models (LMs), but the concrete effect is still unclear. In this paper, we study the effect of representation forgetting on the generality of pre-trained language models, i.e. the potential capability for tackling future downstream tasks. Specifically, we design three metrics, including overall generality destruction (GD), syntactic knowledge forgetting (SynF), and semantic knowledge forgetting (SemF), to measure the evolution of general knowledge in continual learning. With extensive experiments, we find that the generality is destructed in various pre-trained LMs, and syntactic and semantic knowledge is forgotten through continual learning. Based on our experiments and analysis, we further get two insights into alleviating general knowledge forgetting: 1) training on general linguistic tasks at first can mitigate general knowledge forgetting; 2) the hybrid continual learning method can mitigate the generality destruction and maintain more general knowledge compared with those only considering rehearsal or regularization.

Key facts. Cited by 19 papers in the map (data through 7 Oct 2026). Builds on 33 of its 40 references that are in the map. First posted on arXiv 10 May 2023.

Builds on

33 of 40 references.

References · 33

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2020Language Models are Few-Shot Learnerscited by 32,199
  4. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  5. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  6. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  7. 2019DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lightercited by 4,243
  8. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,115
  9. 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,772
  10. 2017A Broad-Coverage Challenge Corpus for Sentence Understanding through Inferencecited by 2,665
  11. 2016Learning without Forgettingcited by 2,465
  12. 2015Character-level Convolutional Networks for Text Classificationcited by 2,293
21 more
  1. 2017Gradient Episodic Memory for Continual Learningcited by 1,871
  2. 2022Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?cited by 1,413
  3. 2021Calibrate Before Use: Improving Few-Shot Performance of Language Modelscited by 1,371
  4. 2020Making Pre-trained Language Models Better Few-shot Learnerscited by 1,255
  5. 2017Memory Aware Synapses: Learning what (not) to forgetcited by 1,094
  6. 2019BERT Rediscovers the Classical NLP Pipelinecited by 1,046
  7. 2018Neural Network Acceptability Judgmentscited by 952
  8. 2020Dark Experience for General Continual Learning: a Strong, Simple Baselinecited by 895
  9. 2022Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learningcited by 886
  10. 2020It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learnerscited by 668
  11. 2018Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interferencecited by 511
  12. 2019What do you learn from context? Probing for sentence structure in contextualized word representationscited by 436
  13. 2018SentEval: An Evaluation Toolkit for Universal Sentence Representationscited by 294
  14. 2020Semantic Drift Compensation for Class-Incremental Learningcited by 241
  15. 2019Episodic Memory in Lifelong Language Learningcited by 188
  16. 2020What Happens To BERT Embeddings During Fine-tuning?cited by 113
  17. 2021Continual Learning for Text Classification with Information Disentanglement Based Regularizationcited by 86
  18. 2022Continual Learning of Natural Language Processing Tasks: A Surveycited by 83
  19. 2022Probing Representation Forgetting in Supervised and Unsupervised Continual Learningcited by 74
  20. 2022Predicting Fine-Tuning Performance with Probingcited by 13
  21. 2022Alleviating Representational Shift for Continual Fine-tuningcited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 2; 2024: 12; 2025: 3; 2026: 2.

19 papers.

Sort:

By area:

  1. 2023An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuningcited by 622
  2. 2023Orthogonal Subspace Learning for Language Model Continual Learningcited by 296
  3. 2024Continual Learning of Large Language Models: A Comprehensive Surveycited by 228
  4. 2024Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMscited by 158
  5. 2024Eight Methods to Evaluate Robust Unlearning in LLMscited by 153
  6. 2024Defending Against Unforeseen Failure Modes with Latent Adversarial Trainingcited by 82
  7. 2024Investigating Continual Pretraining in Large Language Models: Insights and Implicationscited by 54
  8. 2024Localize-and-Stitch: Efficient Model Merging via Sparse Task Arithmeticcited by 50
  9. 2024SAPT: A Shared Attention Framework for Parameter-Efficient Continual Learning of Large Language Modelscited by 37
  10. 2024MIA-Tuner: Adapting Large Language Models as Pre-training Text Detectorcited by 13
  11. 2024PMoE: Progressive Mixture of Experts with Asymmetric Transformer for Continual Learningcited by 8
  12. 2025DiTASK: Multi-Task Fine-Tuning with Diffeomorphic Transformationscited by 7
7 more
  1. 2025Measuring Representational Shifts in Continual Learning: A Linear Transformation Perspectivecited by 5
  2. 2024Balancing Speciality and Versatility: A Coarse to Fine Framework for Mitigating Catastrophic Forgetting in Large Language Modelscited by 3
  3. 2024When Life gives you LLMs, make LLM-ADE: Large Language Models with Adaptive Data Engineeringcited by 1
  4. 2024Efficient Continual Learning for Small Language Models with a Discrete Key-Value Bottleneckcited by 1
  5. 2025On Understanding of the Dynamics of Model Capacity in Continual Learningcited by 1
  6. 2026PEFT-MedSAM: Efficient Fine-Tuning of Medical Foundation Models for Explainable Skin Lesion Segmentationcited by 0
  7. 2026Latent-LoRA: Compact Latent-Space Adapters with Gradient-Free Routing for Continual Learningcited by 0