15 papers
Magnetic Susceptibility of Superconducting Films in the Meissner-London Regime
Ruslan Prozorov
Extracting the London penetration depth, , from the measured magnetic susceptibility, , of a superconducting platelet in a perpendicular magnetic field requires knowing…
A compact six-beam 3D magneto-optical trap based on multifunctional metasurface chips
Chen Qing, Huijie Sun, Yize Liu +5
Magneto-optical traps (MOTs), indispensable to cold-atom-based quantum computing and quantum precision measurement, are increasingly desired in compact and stable form. Particularl…
ConstructCIE: A Dataset for Extracting Causal Information from Construction Accident Narratives
Hung Nguyen, Jaehoon Lee, Namgyun Kim +1
Construction accident narratives contain rich causal information, but the evidence is often implicit, long-span, and distributed. We introduce ConstructCIE, a manually annotated da…
Look Before You Zoom: Adaptive Routing for the Resolution-Context Trade-off in Visual RAG
Oanh N. Tran, Thanh Quoc Hung Le, Oscar Chew +2
Vision-Language Models (VLMs) struggle as query-relevant objects become smaller. To address this, recent training-free approaches dynamically retrieve and zoom into local image reg…
TRACES: Proactive Safety Auditing for Multi-Turn LLM Agents via Trajectory-State Modeling
Jiaqian Li, Yanshu Li, Boxuan Zhang +2
LLM agents increasingly operate through multi-turn tool use and environment interaction, where safety risks often emerge from intermediate steps long before they surface in the fin…
When Correct Demonstrations Hurt: Rethinking the Role of Exemplars in In-Context Learning
Chenghao Qiu, Chunli Peng, Yufeng Yang +2
In-context learning (ICL) is often motivated by the intuition that demonstrations help because they provide correct input-output examples. However, we reveal a counterintuitive phe…
Pop-Up Distractions Reveal Bag-of-Events Behavior in Video Large Language Models
Oscar Chew, Serhii Honcharenko, Qian-Hui Chen +4
A key capability for video understanding is reliably linking subjects to events across time, yet whether Video Large Language Models (VideoLLMs) actually achieve this remains uncle…
The First Drop of Ink: Nonlinear Impact of Distracting Information in Long-Context Reasoning
Muhan Gao, Zih-Ching Chen, Kuan-Hao Huang
As large language models are increasingly deployed in retrieval-augmented generation and agentic systems that accumulate extensive context, understanding how distracting informatio…
Is CLIP Cross-Eyed? Revealing and Mitigating Center Bias in the CLIP Family
Oscar Chew, Hsiao-Ying Huang, Kunal Jain +3
Recent research has shown that contrastive vision-language models such as CLIP often lack fine-grained understanding of visual content. While a growing body of work has sought to a…
Language Steering for Multilingual In-Context Learning
Neeraja Kirtane, Kuan-Hao Huang
If multilingual language model representations contain separable language-associated structure, then shifting between languages may require only simple activation-space interventio…
Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models
Jiaqian Li, Yanshu Li, Kuan-Hao Huang
Steering vectors (SVs) offer a lightweight way to control large language models (LLMs) at inference time by shifting hidden activations, providing a practical middle ground between…
PEPPER: Perception-Guided Perturbation for Robust Backdoor Defense in Text-to-Image Diffusion Models
Oscar Chew, Po-Yi Lu, Jayden Lin +2
Recent studies show that text-to-image (T2I) diffusion models are vulnerable to backdoor attacks, where a trigger in the input prompt can steer generation toward harmful or uninten…
What's Missing in Vision-Language Models? Probing Their Struggles with Causal Order Reasoning
Zhaotian Weng, Haoxuan Li, Xin Eric Wang +2
Despite the impressive performance of vision-language models (VLMs) on downstream tasks, their ability to understand and reason about causal relationships in visual inputs remains…
SHA256 at SemEval-2025 Task 4: Selective Amnesia -- Constrained Unlearning for Large Language Models via Knowledge Isolation
Saransh Agrawal, Kuan-Hao Huang
Large language models (LLMs) frequently memorize sensitive information during training, posing risks when deploying publicly accessible models. Current machine unlearning methods s…
Adaptive Helpfulness-Harmlessness Alignment with Preference Vectors
Ren-Wei Liang, Chin-Ting Hsu, Chan-Hung Yu +6
Ensuring that large language models (LLMs) are both helpful and harmless is a critical challenge, as overly strict constraints can lead to excessive refusals, while permissive mode…