Somesh Jha

15

Papers

101

Total Citations

Papers (15)

AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMs

Validating Mechanistic Interpretations: An Axiomatic Approach

Do Large Code Models Understand Programming Concepts? Counterfactual Analysis for Code Predicates

Two Heads are Actually Better than One: Towards Better Adversarial Robustness via Transduction and Rejection

Robust Attribution Regularization

Attribution-Based Confidence Metric For Deep Neural Networks

A Separation Result Between Data-oblivious and Data-aware Poisoning Attacks

Detecting Errors and Estimating Accuracy on Unlabeled Data with Self-training Ensembles

Overparameterization from Computational Constraints

Robust Learning against Relational Adversaries

A Quantitative Geometric Approach to Neural-Network Smoothness

Grounding Neural Inference with Satisfiability Modulo Theories

Robust and Actively Secure Serverless Collaborative Learning

Analyzing the Robustness of Nearest Neighbors to Adversarial Examples

Reinforcing Adversarial Robustness using Model Confidence Induced by Adversarial Training