13 papers
A reduced model for neoclassical transport in non-axisymmetric magnetic configurations
Y. Ban, X. Garbet, R. Varennes +6
A unified framework is proposed to compute a neoclassical equilibrium for tokamaks with non-axisymmetric magnetic perturbations, based on a Hamiltonian description combined with a…
Charts Are Beyond Pixels: Probing for Layer-Wise Chart Understanding and Editing
Xiaochuan Zhong, Yifan Hou, Chenxi Pang +1
Charts are structured visual compositions whose elements have distinct functional roles, semantic correspondences, and visibility relations. This structural view motivates evaluati…
Synthetic Persona Pretraining: Alignment from Token Zero
Julian Minder, Viktor Moskvoretskii, Raghav Singhal +12
As language-model-based AI is increasingly deployed in autonomous settings, aligning its goals and values with those of humans becomes critical. Today, alignment, and the assistant…
Diversity Matters: Revisiting Test-Time Compute in Vision-Language Models
Yijie Tong, Yifan Hou, Shaobo Cui +2
Test-time compute (TTC) strategies have emerged as a lightweight approach to boost reasoning in large language models (LLMs). However, their application and benefits for vision-lan…
CiteLLM: An Agentic Platform for Trustworthy Scientific Reference Discovery
Mengze Hong, Di Jiang, Chen Jason Zhang +5
Large language models (LLMs) have created new opportunities to enhance the efficiency of scholarly activities; however, challenges persist in the ethical deployment of AI assistanc…
MixAR: Mixture Autoregressive Image Generation
Jinyuan Hu, Jiayou Zhang, Shaobo Cui +2
Autoregressive (AR) approaches, which represent images as sequences of discrete tokens from a finite codebook, have achieved remarkable success in image generation. However, the qu…
Chimera: Diagnosing Shortcut Learning in Visual-Language Understanding
Ziheng Chi, Yifan Hou, Chenxi Pang +3
Diagrams convey symbolic information in a visual format rather than a linear stream of words, making them especially challenging for AI models to process. While recent evaluations…
CoQuIR: A Comprehensive Benchmark for Code Quality-Aware Information Retrieval
Jiahui Geng, Fengyu Cai, Shaobo Cui +8
Code retrieval is essential in modern software development, as it boosts code reuse and accelerates debugging. However, current benchmarks primarily emphasize functional relevance…
Unraveling Misinformation Propagation in LLM Reasoning
Yiyang Feng, Yichen Wang, Shaobo Cui +3
Large Language Models (LLMs) have demonstrated impressive capabilities in reasoning, positioning them as promising tools for supporting human problem-solving. However, what happens…
Nuance Matters: Probing Epistemic Consistency in Causal Reasoning
Shaobo Cui, Junyou Li, Luca Mouchel +2
To address this gap, our study introduces the concept of causal epistemic consistency, which focuses on the self-consistency of Large Language Models (LLMs) in differentiating inte…
A Logical Fallacy-Informed Framework for Argument Generation
Luca Mouchel, Debjit Paul, Shaobo Cui +3
Despite the remarkable performance of Large Language Models (LLMs) in natural language processing tasks, they still struggle with generating logically sound arguments, resulting in…
The Odyssey of Commonsense Causality: From Foundational Benchmarks to Cutting-Edge Reasoning
Shaobo Cui, Zhijing Jin, Bernhard Schölkopf +1
Understanding commonsense causality is a unique mark of intelligence for humans. It helps people understand the principles of the real world better and benefits the decision-making…
Exploring Defeasibility in Causal Reasoning
Shaobo Cui, Lazar Milikic, Yiyang Feng +4
Defeasibility in causal reasoning implies that the causal relationship between cause and effect can be strengthened or weakened. Namely, the causal strength between cause and effec…