الباحثون

Masashi Sugiyama

المنشورات 2

نسخة أولية وصول مفتوح

Preemptive LLM Unlearning against Forbidden Capability Acquisition via Gradient Sealing

Kemou Li, Qizhou Wang, Yue Wang وآخرون · 2026

Open-weight LLMs are released not only as fixed products but also as substrates for downstream fine-tuning. This openness, however, creates legal and ethical risks because users may misuse fine-tuning to instill illicit knowledge or enable hostile operations. Model providers therefore need apre-release defense against …

نسخة أولية وصول مفتوح

Estimating and Orthogonalizing Unknown Pre-training Gradients for Continual Fine-tuning of Large Language Models

Bing Wang, Changchun Li, Xin-Qiang Cai وآخرون · 2026

Continual fine-tuning is essential for large language models (LLMs) to dynamically adapt to real-world environments, yet it inevitably suffers from catastrophic forgetting, particularly the performance degradation of previous tasks and LLMs' general-purpose knowledge. Although existing methods, such as orthogonal gradi …

المؤلفون المشاركون