RLHF and DPO Alignment Effects on Pass@k and Adversarial Robustness in Llama 3 and Mistral

This report synthesises findings from 14 peer-reviewed papers addressing the following research question: How do RLHF and DPO alignment methods differentially impact pass@k scores on HumanEval versus robustness to adversarial code injection attacks in Llama 3 and Mistral models. 9 claims were extracted from source literature; 9 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 9.2/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How do RLHF and DPO alignment methods differentially impact pass@k scores on HumanEval versus robustness to adversarial code injection attacks in Llama 3 and Mistral models?

Autonomous literature synthesis. Automated review score: 9.2/10. Full text and citation available at Assignee Research.