# Uncovering Hidden Violent Tendencies in LLMs: A Demographic Analysis via Behavioral Vignettes

arXiv: https://arxiv.org/abs/2506.20822 · HTML: https://mlsanity.com/paper/2506.20822/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 47 references that are in the map.
- First posted on arXiv 25 Jun 2025.

## Builds on (most cited first, 7 of 7)

- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://mlsanity.com/paper/2006.03654/) (2020)
- [Generative Language Models Exhibit Social Identity Biases](https://mlsanity.com/paper/2310.15819/) (2023)
- [Probing the Emergence of Cross-lingual Alignment during LLM Training](https://mlsanity.com/paper/2406.13229/) (2024)
- [To Protect and To Serve? Analyzing Entity-Centric Framing of Police Violence](https://mlsanity.com/paper/2109.05325/) (2021)
- [Trigger Warnings: Bootstrapping a Violence Detector for FanFiction](https://mlsanity.com/paper/2209.04409/) (2022)

## Built on by (most cited first, 1 of 1)

- [MENTOR: A Metacognition-Driven Self-Evolution Framework for Uncovering and Mitigating Implicit Domain Risks in LLMs](https://mlsanity.com/paper/2511.07107/) (2025)

## Abstract (arXiv)

Large language models (LLMs) are increasingly proposed for detecting and responding to violent content online, yet their ability to reason about morally ambiguous, real-world scenarios remains underexamined. We present the first study to evaluate LLMs using a validated social science instrument designed to measure human response to everyday conflict, namely the Violent Behavior Vignette Questionnaire (VBVQ). To assess potential bias, we introduce persona-based prompting that varies race, age, and geographic identity within the United States. Six LLMs developed across different geopolitical and organizational contexts are evaluated under a unified zero-shot setting. Our study reveals two key findings: (1) LLMs surface-level text generation often diverges from their internal preference for violent responses; (2) their violent tendencies vary across demographics, frequently contradicting established findings in criminology, social science, and psychology.