How to Trick Your AI TA: A Systematic Study of Academic Jailbreaking in LLM Code Evaluation

This article likely presents research on the vulnerabilities of Large Language Models (LLMs) used for code evaluation in academic settings. It investigates methods to bypass the intended constraints and security measures of these AI systems, potentially allowing for unauthorized access or manipulation of the evaluation process. The study's focus on 'jailbreaking' suggests an exploration of techniques to circumvent the AI's safety protocols and achieve unintended outcomes.