2 citations · 3 across the 7 of their papers we have counts for
7 papers
SEVerA: Verified Synthesis of Self-Evolving Agents
Debangshu Banerjee, Changming Xu, Eugene Ie +5
Recent advances have shown the effectiveness of self-evolving LLM agents on tasks such as program repair and scientific discovery. In this paradigm, a planner LLM synthesizes an ag…
Compression Aware Certified Training
Changming Xu, Gagandeep Singh
Deep neural networks deployed in safety-critical, resource-constrained environments must balance efficiency and robustness. Existing methods treat compression and certified robustn…
Robust Answers, Fragile Logic: Probing the Decoupling Hypothesis in LLM Reasoning
Enyi Jiang, Changming Xu, Nischay Singh +2
While Chain-of-Thought (CoT) prompting has become a cornerstone for complex reasoning in Large Language Models (LLMs), the faithfulness of the generated reasoning remains an open q…
Support is All You Need for Certified VAE Training
Changming Xu, Debangshu Banerjee, Deepak Vasisht +1
Variational Autoencoders (VAEs) have become increasingly popular and deployed in safety-critical applications. In such applications, we want to give certified probabilistic guarant…
Cross-Input Certified Training for Universal Perturbations
Changming Xu, Gagandeep Singh
Existing work in trustworthy machine learning primarily focuses on single-input adversarial perturbations. In many real-world attack scenarios, input-agnostic adversarial attacks,…
Bypassing the Safety Training of Open-Source LLMs with Priming Attacks
Jason Vega, Isha Chaudhary, Changming Xu +1
With the recent surge in popularity of LLMs has come an ever-increasing need for LLM safety training. In this paper, we investigate the fragility of SOTA open-source LLMs under sim…
Robust Universal Adversarial Perturbations
Changming Xu, Gagandeep Singh
Universal Adversarial Perturbations (UAPs) are imperceptible, image-agnostic vectors that cause deep neural networks (DNNs) to misclassify inputs with high probability. In practica…