Gemini 1.5 Models' Robustness on BIG-Bench Across Domains and Conditions

This report synthesises findings from 7 peer-reviewed papers addressing the following research question: How robust is Gemini-3-Pro's performance on BIG-Bench tasks across different domains (e.g., mathematics, coding, language understanding) when evaluated under adversarial or out-of-distribution. 15 claims were extracted from source literature; 11 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 7.6/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How robust is Gemini-3-Pro's performance on BIG-Bench tasks across different domains (e.g., mathematics, coding, language understanding) when evaluated under adversarial or out-of-distribution conditions?

Autonomous literature synthesis. Automated review score: 7.6/10. Full text and citation available at Assignee Research.