Scaling Patterns in Adversarial Alignment: Evidence from Multi-LLM Jailbreak Experiments

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nathanson, Samuel, Williams, Rebecca, Matuszek, Cynthia
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915702844162048
author Nathanson, Samuel
Williams, Rebecca
Matuszek, Cynthia
author_facet Nathanson, Samuel
Williams, Rebecca
Matuszek, Cynthia
contents Large language models (LLMs) increasingly operate in multi-agent and safety-critical settings, raising open questions about how their vulnerabilities scale when models interact adversarially. This study examines whether larger models can systematically jailbreak smaller ones - eliciting harmful or restricted behavior despite alignment safeguards. Using standardized adversarial tasks from JailbreakBench, we simulate over 6,000 multi-turn attacker-target exchanges across major LLM families and scales (0.6B-120B parameters), measuring both harm score and refusal behavior as indicators of adversarial potency and alignment integrity. Each interaction is evaluated through aggregated harm and refusal scores assigned by three independent LLM judges, providing a consistent, model-based measure of adversarial outcomes. Aggregating results across prompts, we find a strong and statistically significant correlation between mean harm and the logarithm of the attacker-to-target size ratio (Pearson r = 0.51, p < 0.001; Spearman rho = 0.52, p < 0.001), indicating that relative model size correlates with the likelihood and severity of harmful completions. Mean harm score variance is higher across attackers (0.18) than across targets (0.10), suggesting that attacker-side behavioral diversity contributes more to adversarial outcomes than target susceptibility. Attacker refusal frequency is strongly and negatively correlated with harm (rho = -0.93, p < 0.001), showing that attacker-side alignment mitigates harmful responses. These findings reveal that size asymmetry influences robustness and provide exploratory evidence for adversarial scaling patterns, motivating more controlled investigations into inter-model alignment and safety.
format Preprint
id arxiv_https___arxiv_org_abs_2511_13788
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Scaling Patterns in Adversarial Alignment: Evidence from Multi-LLM Jailbreak Experiments
Nathanson, Samuel
Williams, Rebecca
Matuszek, Cynthia
Machine Learning
Artificial Intelligence
Computation and Language
Cryptography and Security
Multiagent Systems
68T50
I.2.7; I.2.8
Large language models (LLMs) increasingly operate in multi-agent and safety-critical settings, raising open questions about how their vulnerabilities scale when models interact adversarially. This study examines whether larger models can systematically jailbreak smaller ones - eliciting harmful or restricted behavior despite alignment safeguards. Using standardized adversarial tasks from JailbreakBench, we simulate over 6,000 multi-turn attacker-target exchanges across major LLM families and scales (0.6B-120B parameters), measuring both harm score and refusal behavior as indicators of adversarial potency and alignment integrity. Each interaction is evaluated through aggregated harm and refusal scores assigned by three independent LLM judges, providing a consistent, model-based measure of adversarial outcomes. Aggregating results across prompts, we find a strong and statistically significant correlation between mean harm and the logarithm of the attacker-to-target size ratio (Pearson r = 0.51, p < 0.001; Spearman rho = 0.52, p < 0.001), indicating that relative model size correlates with the likelihood and severity of harmful completions. Mean harm score variance is higher across attackers (0.18) than across targets (0.10), suggesting that attacker-side behavioral diversity contributes more to adversarial outcomes than target susceptibility. Attacker refusal frequency is strongly and negatively correlated with harm (rho = -0.93, p < 0.001), showing that attacker-side alignment mitigates harmful responses. These findings reveal that size asymmetry influences robustness and provide exploratory evidence for adversarial scaling patterns, motivating more controlled investigations into inter-model alignment and safety.
title Scaling Patterns in Adversarial Alignment: Evidence from Multi-LLM Jailbreak Experiments
topic Machine Learning
Artificial Intelligence
Computation and Language
Cryptography and Security
Multiagent Systems
68T50
I.2.7; I.2.8
url https://arxiv.org/abs/2511.13788