Fu Xiao
2 papers · Latest:
Cryptography & Security
Safety Anchor: Defending Harmful Fine-tuning via Geometric Bottlenecks
Safety Anchor introduces Safety Bottleneck Regularization (SBR) to defend LLMs against harmful fine-tuning by anchoring hidden states in the unembedding layer.
2605.05995
Cryptography & SecurityMAS-SZZ: Multi-Agentic SZZ Algorithm for Vulnerability-Inducing Commit Identification
MAS-SZZ is a multi-agentic algorithm that accurately identifies vulnerability-inducing commits, significantly outperforming existing SZZ methods.
2604.24398
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.