10 papers
From Nominal Capacity to Deliverable Compute: Design, Calibration and Validation of the D-PIN SCU Dynamic Metering System
Cheng Li, Jiexiong Liu, Shikun Wang +1
Nominal chip specifications do not capture the compute deliverable by heterogeneous distributed networks under communication, runtime and verification constraints. We formalize D-P…
Train What You Deploy:Token-Faithful Post-Training of a Production Coding
Cheng Li, Jiexiong Liu, Yixuan Chen +1
Existing post-training pipelines for coding and terminal agents suffer severe token and control fidelity errors: simplified training environments mismatch production deployments, a…
Efficient and Privacy Aware Edge Cloud Collaborative Inference for Large Language Models
Cheng Li, Jiexiong Liu, Yixuan Chen +1
On-device LLM inference faces a trilemma of response latency, limited hardware resources and user privacy. Full cloud inference delivers strong computing power but exposes user pro…
Oyster-II: Reinforcement Learning for Constructive Safety Alignment in Large Language Models
Jiyang Guan, Yong Xie, Jun Chen +6
Large language models (LLMs) have demonstrated remarkable capabilities across diverse applications, yet ensuring their simultaneous safety, helpfulness, and trustworthiness remains…
A Causal Perspective for Enhancing Jailbreak Attack and Defense
Licheng Pan, Yunsheng Lu, Jiexi Liu +5
Uncovering the mechanisms behind "jailbreaks" in large language models (LLMs) is crucial for enhancing their safety and reliability, yet these mechanisms remain poorly understood.…
Oyster-I: Beyond Refusal -- Constructive Safety Alignment for Responsible Language Models
Ranjie Duan, Jiexi Liu, Xiaojun Jia +27
Large language models (LLMs) typically deploy safety mechanisms to prevent harmful content generation. Most current approaches focus narrowly on risks posed by malicious actors, of…
Strata-Sword: A Hierarchical Safety Evaluation towards LLMs based on Reasoning Complexity of Jailbreak Instructions
Shiji Zhao, Ranjie Duan, Jiexi Liu +10
Large language models (LLMs) have gained widespread recognition for their superior comprehension and have been deployed across numerous domains. Building on Chain-of-Thought (CoT)…
MoAPT: Mixture of Adversarial Prompt Tuning for Vision-Language Models
Shiji Zhao, Qihui Zhu, Shukun Xiong +7
Large pre-trained Vision Language Models (VLMs) demonstrate excellent generalization capabilities but remain highly susceptible to adversarial examples, posing potential security r…
Inverse Reinforcement Learning with Dynamic Reward Scaling for LLM Alignment
Ruoxi Cheng, Haoxuan Ma, Weixin Wang +7
Alignment is vital for safely deploying large language models (LLMs). Existing techniques are either reward-based (training a reward model on preference pairs and optimizing with r…
Mirage in the Eyes: Hallucination Attack on Multi-modal Large Language Models with Only Attention Sink
Yining Wang, Mi Zhang, Junjie Sun +6
Fusing visual understanding into language generation, Multi-modal Large Language Models (MLLMs) are revolutionizing visual-language applications. Yet, these models are often plague…