@inproceedings{sun-etal-2025-unveiling,
    title = "Unveiling and Addressing Pseudo Forgetting in Large Language Models",
    author = "Sun, Huashan  and
      Yang, Yizhe  and
      Li, Yinghao  and
      Li, Jiawei  and
      Gao, Yang",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Findings of the Association for Computational Linguistics: ACL 2025",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.findings-acl.1212/",
    pages = "23642--23658",
    ISBN = "979-8-89176-256-5",
    abstract = "Although substantial efforts have been made to mitigate catastrophic forgetting in continual learning, the intrinsic mechanisms are not well understood. In this work, we demonstrate the existence of ``pseudo forgetting'': the performance degradation in previous tasks is not attributed to a loss of capabilities, but rather to the failure of the instructions to activate the appropriate model capabilities. We show that the model{'}s performance on previous tasks can be restored through two simple interventions: (1) providing partial external correct rationale, and (2) appending semantically meaningless suffixes to the original instructions, to guide the generation of correct rationales. Through empirical analysis of the internal mechanisms governing rationale generation, we reveal that models exhibiting pseudo forgetting show reduced instruction dependence during rationale generation, leading to suboptimal activation of their inherent capabilities. Based on this insight, we propose Rationale-Guidance Difficulty based Replay (RGD-R) framework that dynamically allocates replay data based on the model{'}s ability to correctly leverage the intrinsic capabilities. Experimental results demonstrate that RGD-R effectively mitigates pseudo forgetting while maintaining model plasticity."
}