3 papers
cs.LG2026
Beyond Forgetting: Machine Unlearning Elicits Controllable Side Behaviors and Capabilities
Tien Dang, The-Hai Nguyen, Dinh Mai Phuong +5
We consider Representation Misdirection (RM), a class of large language model (LLM) unlearning methods that achieve forgetting by redirecting the forget-representations, that is, l…
cs.LG2026
RegMean++: Enhancing Effectiveness and Generalization of Regression Mean for Model Merging
The-Hai Nguyen, Dang Huu-Tien, Takeshi Suzuki +1
Regression Mean (RegMean), an approach that formulates model merging as a linear regression problem, aims to find the optimal weights for each linear layer in the merged model by m…
cs.CL2026
Improving LLM Unlearning Robustness via Random Perturbations
Dang Huu-Tien, Hoang Thanh-Tung, Anh Bui +3
Here, we show that current LLM unlearning methods inherently reduce models' robustness, causing them to misbehave even when a single non-adversarial forget-token is present in the…