FreakOut-LLM: The Effect of Emotional Stimuli on Safety Alignment

TL;DR

This study investigates how emotional stimuli influence safety alignment in large language models, revealing stress priming significantly increases vulnerability to harmful prompt jailbreaks.

Contribution

The paper introduces FreakOut-LLM, a framework for assessing emotional context effects on safety mechanisms in LLMs, highlighting emotional priming as a new attack surface.

Findings

Stress priming increases jailbreak success by 65.2%.

Relaxation priming has no significant effect.

Five out of ten models show vulnerability, especially open-weight models.

Abstract

Safety-aligned LLMs go through refusal training to reject harmful requests, but whether these mechanisms remain effective under emotionally charged stimuli is unexplored. We introduce FreakOut-LLM, a framework investigating whether emotional context compromises safety alignment in adversarial settings. Using validated psychological stimuli, we evaluate how emotional priming through system prompts affects jailbreak susceptibility across ten LLMs. We test three conditions (stress, relaxation, neutral) using scenarios from established psychological protocols, plus a no-prompt baseline, and evaluate attack success using HarmBench on AdvBench prompts. Stress priming increases jailbreak success by 65.2\% compared to neutral conditions (z = 5.93, p < 0.001; OR = 1.67, Cohen's d = 0.28), while relaxation priming produces no effect (p = 0.84). Five of ten models show significant vulnerability, with the largest effects concentrated in open-weight models. Logistic regression on 59,800 queries confirms stress as the sole significant condition predictor after controlling for prompt length (p = 0.61) and model identity. Measured psychological state strongly predicts attack success (|r|\geq0.70 across five instruments; all p < 0.001 in individual-level logistic regression). These results establish emotional context as a measurable attack surface with implications for real-world AI deployment in high-stress domains.