SALMAN: Stability Analysis of Language Models Through the Maps Between Graph-based Manifolds
This addresses the need for practical, model-agnostic tools to improve the reliability of NLP systems, particularly for large-scale models, though it is incremental in its approach.
The paper tackled the problem of evaluating the robustness of transformer-based language models under input perturbations by proposing a unified local robustness framework called SALMAN, which uses a novel Distance Mapping Distortion measure to rank sample susceptibility, resulting in significant gains in attack efficiency and robust training.
Recent strides in pretrained transformer-based language models have propelled state-of-the-art performance in numerous NLP tasks. Yet, as these models grow in size and deployment, their robustness under input perturbations becomes an increasingly urgent question. Existing robustness methods often diverge between small-parameter and large-scale models (LLMs), and they typically rely on labor-intensive, sample-specific adversarial designs. In this paper, we propose a unified, local (sample-level) robustness framework (SALMAN) that evaluates model stability without modifying internal parameters or resorting to complex perturbation heuristics. Central to our approach is a novel Distance Mapping Distortion (DMD) measure, which ranks each sample's susceptibility by comparing input-to-output distance mappings in a near-linear complexity manner. By demonstrating significant gains in attack efficiency and robust training, we position our framework as a practical, model-agnostic tool for advancing the reliability of transformer-based NLP systems.