Paper-Conference

Correlating Adversarial Robustness with the Energy of Untrained Classifiers
Neural networks are known to be vulnerable to imperceptible perturbations to their inputs, causing unwarranted behavior in the …
Cite
Correlating Adversarial Robustness with the Energy of Untrained Classifiers
Neutralizing Proactive Defense using Diffusion-based Upsampling
The rapid spread of open-source generative models has made it easy to create highly realistic manipulated media, posing a critical …
PDF Cite DOI
Neutralizing Proactive Defense using Diffusion-based Upsampling
Steering Vectors are an Adversarial Attack Surface
Activation steering has become a popular way to control Large Language Model (LLM) behavior without fine-tuning. Since the technique is …
PDF Cite Code
Steering Vectors are an Adversarial Attack Surface
Multi-objective Evolutionary Merging Enables Efficient Reasoning Models
Reasoning models have demonstrated remarkable capabilities in solving complex problems by leveraging long chains of thought. However, …
PDF Cite
Multi-objective Evolutionary Merging Enables Efficient Reasoning Models
A Provable Energy-Guided Test-Time Defense Boosting Adversarial Robustness of Large Vision-Language Models
Despite the rapid progress in multimodal models and Large Visual-Language Models (LVLM), they remain highly susceptible to adversarial …
PDF Cite Code
A Provable Energy-Guided Test-Time Defense Boosting Adversarial Robustness of Large Vision-Language Models
Harnessing Hyperbolic Geometry for Harmful Prompt Detection and Sanitization
Vision–Language Models (VLMs) have become essential for tasks such as image synthesis, captioning, and retrieval by aligning textual …
PDF Cite
Harnessing Hyperbolic Geometry for Harmful Prompt Detection and Sanitization
Implicit Inversion turns CLIP into a Decoder
CLIP is a discriminative model trained to align images and text in a shared embedding space. Due to its multimodal structure, it serves …
PDF Cite Code
Implicit Inversion turns CLIP into a Decoder
Spilled Energy in Large Language Models
We reinterpret the final softmax classifier over the vocabulary of Large Language Models (LLM) as an Energy-based Model (EBM). This …
Cite
Spilled Energy in Large Language Models
Why Adversarially Train Diffusion Models?
We answer the question in the title showing that adversarial training (AT) for diffusion models (DMs) fundamentally differs from …
PDF Cite Code
Why Adversarially Train Diffusion Models?
MASS : MoErging through Adaptive Subspace Selection
Model merging has recently emerged as a lightweight alternative to ensembling, combining multiple fine-tuned models into a single set …
PDF Cite Code
MASS : MoErging through Adaptive Subspace Selection