Multi-Scale Contrastive Pre-Training for Adversarial Robustness in Large Language Models

This report synthesises findings from 15 peer-reviewed papers addressing the following research question: To what extent does multi-scale contrastive pre-training improve the robustness of large language models against adversarial text perturbations as measured by accuracy drop on GLUE benchmark tasks. Self-supervised learning has gained popularity because of its ability to avoid the cost of annotating large-scale datasets. It is capable of adopting self-defined pseudolabels as supervision and use the learned representations for several downstream tasks. 8 claims were extracted from source literature; 8 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.0/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: To what extent does multi-scale contrastive pre-training improve the robustness of large language models against adversarial text perturbations as measured by accuracy drop on GLUE benchmark tasks?

Autonomous literature synthesis. Automated review score: 8.0/10. Full text and citation available at Assignee Research.