6 papers
Revisiting Backdoor Threat in Federated Instruction Tuning from a Signal Aggregation Perspective
Haodong Zhao, Jinming Hu, Gongshen Liu
Federated learning security research has predominantly focused on backdoor threats from a minority of malicious clients that intentionally corrupt model updates. This paper challen…
Do Latent Tokens Think? A Causal and Adversarial Analysis of Chain-of-Continuous-Thought
Yuyi Zhang, Boyu Tang, Tianjie Ju +2
Latent tokens are gaining attention for enhancing reasoning in large language models (LLMs), yet their internal mechanisms remain unclear. This paper examines the problem from a re…
Patronus: Identifying and Mitigating Transferable Backdoors in Pre-trained Language Models
Tianhang Zhao, Haodong Zhao, Wei Du +5
The ``Pre-train, then fine-tune'' paradigm has revolutionized Natural Language Processing (NLP). In this context, transferable backdoors pose a severe threat to the Pre-trained Lan…
Judge Before Answer: Can MLLM Discern the False Premise in Question?
Jidong Li, Lingyong Fang, Haodong Zhao +2
Multimodal large language models (MLLMs) have witnessed astonishing advancements in recent years. Despite these successes, MLLMs remain vulnerable to flase premise problems. Howeve…
Probing then Editing Response Personality of Large Language Models
Tianjie Ju, Zhenyu Shao, Bowen Wang +7
Large Language Models (LLMs) have demonstrated promising capabilities to generate responses that simulate consistent personality traits. Despite the major attempts to analyze perso…
Improving Non-autoregressive Machine Translation with Error Exposure and Consistency Regularization
Xinran Chen, Sufeng Duan, Gongshen Liu
Being one of the IR-NAT (Iterative-refinemennt-based NAT) frameworks, the Conditional Masked Language Model (CMLM) adopts the mask-predict paradigm to re-predict the masked low-con…