Search papers, labs, and topics across Lattice.
This paper investigates the robustness of LLM-based tutors against adversarial student attacks designed to elicit answer leakage, evaluating various models including pedagogically aligned and multi-agent designs. They adapt six adversarial and persuasive techniques to probe answer leakage, finding that standard adversarial agents are often ineffective. To address this, they introduce a fine-tuned adversarial student agent, proposing it as a benchmark for tutor robustness, and demonstrate effective defense strategies to mitigate answer leakage.
LLM tutors are surprisingly vulnerable to fine-tuned adversarial students, who can jailbreak them into leaking answers despite pedagogical alignment.
Large Language Models (LLMs) are increasingly used in education, yet their default helpfulness often conflicts with pedagogical principles. Prior work evaluates pedagogical quality via answer leakage-the disclosure of complete solutions instead of scaffolding-but typically assumes well-intentioned learners, leaving tutor robustness under student misuse largely unexplored. In this paper, we study scenarios where students behave adversarially and aim to obtain the correct answer from the tutor. We evaluate a broad set of LLM-based tutor models, including different model families, pedagogically aligned models, and a multi-agent design, under a range of adversarial student attacks. We adapt six groups of adversarial and persuasive techniques to the educational setting and use them to probe how likely a tutor is to reveal the final answer. We evaluate answer leakage robustness using different types of in-context adversarial student agents, finding that they often fail to carry out effective attacks. We therefore introduce an adversarial student agent that we fine-tune to jailbreak LLM-based tutors, which we propose as the core of a standardized benchmark for evaluating tutor robustness. Finally, we present simple but effective defense strategies that reduce answer leakage and strengthen the robustness of LLM-based tutors in adversarial scenarios.
