9 papers
Rethinking Multi-Image Re-Representation in Multi-Image Understanding
Gengyuan Zhang, Xiao Han, Xinyu Xie +2
Multi-image understanding requires MLLMs not only to recognise the content of individual images, but also to organise visual evidence distributed across them. We study this problem…
Deep Research Agents Brings Deeper Harm
Shuo Chen, Zonggen Li, Xingyu Jin +8
We reveal that Deep Research (DR) agents systematically expose safety risks: simply submitting harmful queries that a standalone LLM would reject outright can elicit detailed and d…
Does Machine Unlearning Truly Remove Knowledge?
Haokun Chen, Yueqi Zhang, Yuan Bi +9
In recent years, Large Language Models (LLMs) have achieved remarkable advancements, drawing significant attention from the research community. Their capabilities are largely attri…
Memory Helps, but Confabulation Misleads: Understanding Streaming Events in Videos with MLLMs
Gengyuan Zhang, Mingcong Ding, Tong Liu +2
Multimodal large language models (MLLMs) have demonstrated strong performance in understanding videos holistically, yet their ability to process streaming videos-videos are treated…
FocalPO: Enhancing Preference Optimizing by Focusing on Correct Preference Rankings
Tong Liu, Xiao Yu, Wenxuan Zhou +2
Efficient preference optimization algorithms such as Direct Preference Optimization (DPO) have become a popular approach in aligning large language models (LLMs) with human prefere…
Multimodal Pragmatic Jailbreak on Text-to-image Models
Tong Liu, Zhixin Lai, Jiawen Wang +6
Diffusion models have recently achieved remarkable advancements in terms of image quality and fidelity to textual prompts. Concurrently, the safety of such generative models has be…
HiFT: A Hierarchical Full Parameter Fine-Tuning Strategy
Yongkang Liu, Yiqun Zhang, Qian Li +5
Full-parameter fine-tuning has become the go-to choice for adapting language models (LMs) to downstream tasks due to its excellent performance. As LMs grow in size, fine-tuning the…
Temperature-scaling surprisal estimates improve fit to human reading times -- but does it do so for the "right reasons"?
Tong Liu, Iza Škrjanec, Vera Demberg
A wide body of evidence shows that human language processing difficulty is predicted by the information-theoretic measure surprisal, a word's negative log probability in context. H…
On Calibration of Graph Neural Networks for Node Classification
Tong Liu, Yushan Liu, Marcel Hildebrandt +3
Graphs can model real-world, complex systems by representing entities and their interactions in terms of nodes and edges. To better exploit the graph structure, graph neural networ…