Code-to-Natural-Language Pretraining Ratios and CodeT5 Zero-Shot Accuracy on MBPP

This report synthesises findings from 15 peer-reviewed papers addressing the following research question: How does varying the ratio of code-to-natural-language pretraining data affect CodeT5's zero-shot accuracy on the MBPP dataset for low-resource programming languages. 10 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 7.7/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does varying the ratio of code-to-natural-language pretraining data affect CodeT5's zero-shot accuracy on the MBPP dataset for low-resource programming languages?

Autonomous literature synthesis. Automated review score: 7.7/10. Full text and citation available at Assignee Research.