63 citations · 235 across the 69 of their papers we have counts for
8 papers · 1 filter
Conflict-Aware Logit Adapters for Utility-Preserving Anti-Distillation
Quoc Phong Dao, Binh Anh Nguyen, Linh Ngo Van +3
While reasoning traces improve model capability, they also provide supervision signals for downstream distillation. Initial efforts to mitigate this issue rely on decoding-time int…
VGGA: Geometry-Grounded Autoencoder for Video Generation
Chaofan Gan, Yian Zhao, Shijie Li +4
Autoencoders are the core component of latent video generative models, compressing high-dimensional videos into compact latent spaces for downstream generation. However, existing v…
Open Vocabulary Domain Unlearning
Sumanth Udupa, Mehrtash Harandi, Yadan Luo +1
Vision-Language Models (VLMs) exhibit remarkable zero-shot generalization, yet they often encode unwanted or hazardous stylistic domains such as idealized textbook diagrams in medi…
Reliability-Prioritized Fine-Grained Generation in Multimodal Large
Xiaomeng Fan, Wei Wu, Yuwei Wu +9
Multimodal large language models (MLLMs) are increasingly expected to generate fine-grained descriptions of visual content. However, we observe and theoretically show that generati…
Awakening Diffusion Transformers: Eliciting Stronger Generation and Understanding via Massive Activation Modulation
Chaofan Gan, Zicheng Zhao, Yuanpeng Tu +6
Massive Activations (MAs) have been widely observed in Transformer-based models, yet their structure and functional roles in Diffusion Transformers (DiTs) remain insufficiently und…
Test-Time Instance-Specific Parameter Composition: A New Paradigm for Adaptive Generative Modeling
Minh-Tuan Tran, Xuan-May Le, Quan Hung Tran +3
Existing generative models, such as diffusion and auto-regressive networks, are inherently static, relying on a fixed set of pretrained parameters to handle all inputs. In contrast…
Sharpness-Aware Minimization in Logit Space Efficiently Enhances Direct Preference Optimization
Haocheng Luo, Zehang Deng, Thanh-Toan Do +3
Direct Preference Optimization (DPO) has emerged as a popular algorithm for aligning pretrained large language models with human preferences, owing to its simplicity and training s…
Antibody: Strengthening Defense Against Harmful Fine-Tuning for Large Language Models via Attenuating Harmful Gradient Influence
Quoc Minh Nguyen, Trung Le, Jing Wu +2
Fine-tuning-as-a-service introduces a threat to Large Language Models' safety when service providers fine-tune their models on poisoned user-submitted datasets, a process known as…