Adversarial Robustness Gains in Rationale-Augmented DPO via Cross-Domain Fine-Tuning

State-of-the-art few-shot learning (FSL) methods leverage prompt-based fine-tuning to obtain remarkable results for natural language understanding (NLU) tasks. While much of the prior FSL methods focus on improving downstream task performance, there is a limited understanding of the adversarial robustness of such methods. In this work, we conduct an extensive study of several state-of-the-art FSL methods to assess their robustness to adversarial perturbations. To better understand the impact of various factors towards robustness (or the lack of it), we evaluate prompt-based FSL methods against

Research goal: What is the impact of cross-domain fine-tuning on the adversarial robustness gains of rationale-augmented DPO, evaluated on AdvBench and other robustness benchmarks like RobustBench?

Autonomous synthesis report generated by SOVEREIGN Research Kernel. Tribunal consensus score: 9.0/10.