paper-with-me

Papers

Beyond Forgetting: Machine Unlearning Elicits Controllable Side Behaviors and Capabilities

2026-01-29 · Tien Dang, The-Hai Nguyen, Dinh Mai Phuong, Nguyen Minh Phuong, Anh Bui, Hoang Thanh-Tung, Le-Minh Nguyen, Naoya Inoue arxiv

We consider Representation Misdirection (RM), a class of large language model (LLM) unlearning methods that achieve forgetting by redirecting the forget-representations, that is, latent representations of forget-samples, toward a target vector. Despite being important, the roles of the target vector used in RM, however, remain underexplored. Here, we approach and revisit RM through the lens of the Linear Representation Hypothesis. Specifically, if one can identify a one-dimensional representation corresponding to a high-level concept, the Linear Representation Hypothesis enables linear operations on this concept vector within the forget-representation space. Under this view, we hypothesize that, beyond forgetting, machine unlearning via RM elicits controllable emergent side behaviors and stronger side capabilities corresponding to the high-level concept. Our hypothesis is empirically validated across a wide range of tasks, including behavioral control (e.g., controlling unlearned models' truthfulness, sentiment, refusal, and language) and capability enhancement (e.g., improving unlearned models' in-context learning (ICL) capability). Our findings reveal that this phenomenon could be either a hidden risk if misused or a mechanism that can be harnessed for developing unlearned models that require stronger capabilities and controllable behaviors.

📄 PDF Abstract BibTeX arXiv:2601.21702

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

iShumei-Chinchunmei at SemEval-2025 Task 4: A balanced forgetting and retention multi-task framework using effective unlearning loss

2025-07-22 · Yujian Sun, Tian Li arxiv

As the Large Language Model (LLM) gains widespread adoption, increasing attention has been given to the challenge of making LLM forget non-compliant data memorized during its pre-training. Machine Unlearning focuses on e…

LEGATO: Good Identity Unlearning Is Continuous

2026-01-07 · Qiang Chen, Chun-Wun Cheng, Xiu Su, Hongyan Xu 외 arxiv

Machine unlearning has become a crucial role in enabling generative models trained on large datasets to remove sensitive, private, or copyright-protected data. However, existing machine unlearning methods face three chal…

Controllable Machine Unlearning via Gradient Pivoting

2025-10-22 · Youngsik Hwang, Dong-Young Lim arxiv

Machine unlearning (MU) aims to remove the influence of specific data from a trained model. However, approximate unlearning methods, often formulated as a single-objective optimization (SOO) problem, face a critical trad…

Neural Machine Unranking

2024-08-09 · Jingrui Hou, Axel Finke, Georgina Cosma

We tackle the problem of machine unlearning within neural information retrieval, termed Neural Machine UnRanking (NuMuR) for short. Many of the mainstream task- or model-agnostic approaches for machine unlearning were de…

Information RetrievalMachine UnlearningRetrieval

FROC: A Unified Framework with Risk-Optimized Control for Machine Unlearning in LLMs

2025-12-15 · Si Qi Goh, Yongsen Zheng, Ziyao Liu, Sami Hormi 외 arxiv

Machine unlearning (MU) seeks to eliminate the influence of specific training examples from deployed models. As large language models (LLMs) become widely used, managing risks arising from insufficient forgetting or util…