1 citations · 2 across the 7 of their papers we have counts for
8 papers
Towards Robustness against Typographic Attack with Training-free Concept Localization
Bohan Liu, Wenqian Ye, Guangzhi Xiong +3
Models trained via Contrastive Language-Image Pretraining (CLIP) serve as the foundational vision encoders for most modern Large Vision Language Models (LVLMs). Despite their wides…
Attention-guided Fine-tuning of Multimodal Large Language Models Improves Chain-of-Thought Reasoning
Sanchit Sinha, Guangzhi Xiong, Bohan Liu +2
The effectiveness of Chain-of-Thought (CoT) prompting in Multimodal Large Language Models (MLLMs) remains uncertain: across several visual reasoning benchmarks, CoT prompting often…
Retrieving Counterfactuals Improves Visual In-Context Learning
Guangzhi Xiong, Sanchit Sinha, Zhenghao He +1
Vision-language models (VLMs) have achieved impressive performance across a wide range of multimodal reasoning tasks, but they often struggle to disentangle fine-grained visual att…
CASL: Concept-Aligned Sparse Latents for Interpreting Diffusion Models
Zhenghao He, Guangzhi Xiong, Boyang Wang +2
Internal activations of diffusion models encode rich semantic information, but interpreting such representations remains challenging. While Sparse Autoencoders (SAEs) have shown pr…
Triggering Chain-of-Thought via Latent Feature Interventions in Large Language Models
Zhenghao He, Guangzhi Xiong, Bohan Liu +2
Chain-of-Thought (CoT) prompting often improves the reasoning performance of large language models (LLMs), but the internal signal that triggers this behavior remains poorly unders…
Toward Faithful Retrieval-Augmented Generation with Sparse Autoencoders
Guangzhi Xiong, Zhenghao He, Bohan Liu +2
Retrieval-Augmented Generation (RAG) improves the factuality of large language models (LLMs) by grounding outputs in retrieved evidence, but faithfulness failures, where generation…
Concept-RuleNet: Grounded Multi-Agent Neurosymbolic Reasoning in Vision Language Models
Sanchit Sinha, Guangzhi Xiong, Zhenghao He +1
Modern vision-language models (VLMs) deliver impressive predictive accuracy yet offer little insight into 'why' a decision is reached, frequently hallucinating facts, particularly…
GCAV: A Global Concept Activation Vector Framework for Cross-Layer Consistency in Interpretability
Zhenghao He, Sanchit Sinha, Guangzhi Xiong +1
Concept Activation Vectors (CAVs) provide a powerful approach for interpreting deep neural networks by quantifying their sensitivity to human-defined concepts. However, when comput…