Linked Presentation: Instruction Backdoor Attacks Against Customized LLMsScalable Zero-knowledge Proofs for Non-linear Functions in Machine Learning
