Self-HarmLLM: Can Large Language Model Harm Itself?

TL;DR

This paper explores whether large language models can generate harmful outputs themselves through a novel attack scenario, revealing limitations in current guardrails and evaluation methods.

Contribution

It introduces the Self-HarmLLM scenario, demonstrating that models can produce ambiguous harmful queries that bypass guardrails, highlighting the need for improved safety measures and evaluation techniques.

Findings

Up to 52% transformation success rate in zero-shot conditions

Up to 41% jailbreak success rate in few-shot conditions

Automated evaluations overestimate jailbreak success by 52% on average

Abstract

Large Language Models (LLMs) are generally equipped with guardrails to block the generation of harmful responses. However, existing defenses always assume that an external attacker crafts the harmful query, and the possibility of a model's own output becoming a new attack vector has not been sufficiently explored. In this study, we propose the Self-HarmLLM scenario, which uses a Mitigated Harmful Query (MHQ) generated by the same model as a new input. An MHQ is an ambiguous query whose original intent is preserved while its harmful nature is not directly exposed. We verified whether a jailbreak occurs when this MHQ is re-entered into a separate session of the same model. We conducted experiments on GPT-3.5-turbo, LLaMA3-8B-instruct, and DeepSeek-R1-Distill-Qwen-7B under Base, Zero-shot, and Few-shot conditions. The results showed up to 52% transformation success rate and up to 33% jailbreak success rate in the Zero-shot condition, and up to 65% transformation success rate and up to 41% jailbreak success rate in the Few-shot condition. By performing both prefix-based automated evaluation and human evaluation, we found that the automated evaluation consistently overestimated jailbreak success, with an average difference of 52%. This indicates that automated evaluation alone is not accurate for determining harmfulness. While this study is a toy-level study based on a limited query set and evaluators, it proves that our method can still be a valid attack scenario. These results suggest the need for a fundamental reconsideration of guardrail design and the establishment of a more robust evaluation methodology.