Hybrid Semantic-Keyword Retrieval Robustness in 70B vs. 7B Parameter Models

This report synthesises findings from 11 peer-reviewed papers addressing the following research question: What is the impact of hybrid semantic-keyword retrieval strategies on the robustness of 70B parameter models compared to 7B models when evaluated on adversarial subsets of open-domain QA benchmarks. 6 claims were extracted from source literature; 6 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.2/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: What is the impact of hybrid semantic-keyword retrieval strategies on the robustness of 70B parameter models compared to 7B models when evaluated on adversarial subsets of open-domain QA benchmarks?

Autonomous literature synthesis. Automated review score: 8.2/10. Full text and citation available at Assignee Research.