Manifold-Aware Cross-Encoders Outperform Dense Retrievers on Adversarial Benchmarks

This report synthesises findings from 13 peer-reviewed papers addressing the following research question: Can cross-encoder models trained with manifold-aware objectives outperform traditional dense retrievers on adversarial benchmarks like Adversarial NQ while maintaining competitive accuracy on. 10 claims were extracted from source literature; 10 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 9.0/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: Can cross-encoder models trained with manifold-aware objectives outperform traditional dense retrievers on adversarial benchmarks like Adversarial NQ while maintaining competitive accuracy on standard retrieval tasks?

Autonomous literature synthesis. Automated review score: 9.0/10. Full text and citation available at Assignee Research.