11 citations · 15 across the 15 of their papers we have counts for
17 papers
Restoring General and Improving Domain-Specific Safety in Fine-Tuned LLMs Without Retraining
Vishnu Kabir Chhabra, Ding Zhu, Zhihui Zhu +2
Safety-aligned language models are known to suffer degradation in refusal behavior following domain-specific fine-tuning, even when the fine-tuning data is entirely benign. The pre…
One Token Can Be Enough: Bridging Prompting and Activation Steering with Prefix Steering
Xudong Zhu, Zhihui Zhu
Prompting guides language model behavior through the initial context, whereas activation steering often intervenes throughout generation. A natural question is whether steering can…
SMCEvolve: Principled Scientific Discovery via Sequential Monte Carlo Evolution
Jiachen Jiang, Huminhao Zhu, Zhihui Zhu
LLM-driven program evolution has emerged as a powerful tool for automated scientific discovery, yet existing frameworks offer no principled guide for designing their individual com…
DeltaEvolve: Accelerating Scientific Discovery through Momentum-Driven Evolution
Jiachen Jiang, Tianyu Ding, Zhihui Zhu
LLM-driven evolutionary systems have shown promise for automated science discovery, yet existing approaches such as AlphaEvolve rely on full-code histories that are context-ineffic…
Improving Visual Discriminability of CLIP for Training-Free Open-Vocabulary Semantic Segmentation
Jinxin Zhou, Jiachen Jiang, Zhihui Zhu
Extending CLIP models to semantic segmentation remains challenging due to the misalignment between their image-level pre-training objectives and the pixel-level visual understandin…
From Emergence to Control: Probing and Modulating Self-Reflection in Language Models
Xudong Zhu, Jiachen Jiang, Mohammad Mahdi Khalili +1
Self-reflection -- the ability of a large language model (LLM) to revisit, evaluate, and revise its own reasoning -- has recently emerged as a powerful behavior enabled by reinforc…
Understanding Task Vectors in In-Context Learning: Emergence, Functionality, and Limitations
Yuxin Dong, Jiachen Jiang, Zhihui Zhu +1
Task vectors offer a compelling mechanism for accelerating inference in in-context learning (ICL) by distilling task-specific information into a single, reusable representation. De…
Analyzing Fine-Grained Alignment and Enhancing Vision Understanding in Multimodal Language Models
Jiachen Jiang, Jinxin Zhou, Bo Peng +2
Achieving better alignment between vision embeddings and Large Language Models (LLMs) is crucial for enhancing the abilities of Multimodal LLMs (MLLMs), particularly for recent mod…
From Compression to Expression: A Layerwise Analysis of In-Context Learning
Jiachen Jiang, Yuxin Dong, Jinxin Zhou +1
In-context learning (ICL) enables large language models (LLMs) to adapt to new tasks without weight updates by learning from demonstration sequences. While ICL shows strong empiric…
Learning Where and What to Restore for Composite Image Restoration
Jiachen Jiang, Tianyu Ding, Ke Zhang +5
Real-world degraded images often contain multiple co-occurring degradation types, making composite image restoration fundamentally different from the single-degradation setting ass…
Analyzing and Mitigating Model Collapse in Rectified Flow Models
Huminhao Zhu, Fangyikang Wang, Tianyu Ding +2
Training with synthetic data is becoming increasingly inevitable as synthetic content proliferates across the web, driven by the remarkable performance of recent deep generative mo…
Captions Speak Louder than Images: Generalizing Foundation Models for E-commerce from High-quality Multimodal Instruction Data
Xinyi Ling, Hanwen Du, Bo Peng +2
Leveraging multimodal data to drive breakthroughs in e-commerce applications through Multimodal Foundation Models (MFMs) is gaining increasing attention from the research community…
Tracing Representation Progression: Analyzing and Enhancing Layer-Wise Similarity
Jiachen Jiang, Jinxin Zhou, Zhihui Zhu
Analyzing the similarity of internal representations has been an important technique for understanding the behavior of deep neural networks. Most existing methods for analyzing the…
AdaContour: Adaptive Contour Descriptor with Hierarchical Representation
Tianyu Ding, Jinxin Zhou, Tianyi Chen +3
Existing angle-based contour descriptors suffer from lossy representation for non-starconvex shapes. By and large, this is the result of the shape being registered with a single gl…
OTOv3: Automatic Architecture-Agnostic Neural Network Training and Compression from Structured Pruning to Erasing Operators
Tianyi Chen, Tianyu Ding, Zhihui Zhu +4
Compressing a predefined deep neural network (DNN) into a compact sub-network with competitive performance is crucial in the efficient machine learning realm. This topic spans vari…
The Efficiency Spectrum of Large Language Models: An Algorithmic Survey
Tianyu Ding, Tianyi Chen, Haidong Zhu +7
The rapid growth of Large Language Models (LLMs) has been a driving force in transforming various domains, reshaping the artificial general intelligence landscape. However, the inc…
DREAM: Diffusion Rectification and Estimation-Adaptive Models
Jinxin Zhou, Tianyu Ding, Tianyi Chen +4
We present DREAM, a novel training framework representing Diffusion Rectification and Estimation Adaptive Models, requiring minimal code changes (just three lines) yet significantl…