AREG: Adversarial Resource Extraction Game for Evaluating Persuasion and Resistance in Large Language Models

TL;DR

The paper introduces AREG, a benchmark for assessing both persuasion and resistance in large language models through adversarial negotiations, revealing their weak correlation and systematic resistance advantage.

Contribution

It presents a novel multi-turn negotiation framework to evaluate social influence in LLMs, highlighting the dissociation between persuasion and resistance capabilities.

Findings

Resistance scores are higher than persuasion scores across models.

Weak correlation ($\rho=0.33$) between persuasion and resistance.

Interaction structure significantly influences social influence outcomes.

Abstract

Evaluating the social intelligence of Large Language Models (LLMs) increasingly requires moving beyond static text generation toward dynamic, adversarial interaction. We introduce the Adversarial Resource Extraction Game (AREG), a benchmark that operationalizes persuasion and resistance as a multi-turn, zero-sum negotiation over financial resources. Using a round-robin tournament across frontier models, AREG enables joint evaluation of offensive (persuasion) and defensive (resistance) capabilities within a single interactional framework. Our analysis provides evidence that these capabilities are weakly correlated ($\rho = 0.33$) and empirically dissociated: strong persuasive performance does not reliably predict strong resistance, and vice versa. Across all evaluated models, resistance scores exceed persuasion scores, indicating a systematic defensive advantage in adversarial dialogue settings. Further linguistic analysis suggests that interaction structure plays a central role in these outcomes. Incremental commitment-seeking strategies are associated with higher extraction success, while verification-seeking responses are more prevalent in successful defenses than explicit refusal. Together, these findings indicate that social influence in LLMs is not a monolithic capability and that evaluation frameworks focusing on persuasion alone may overlook asymmetric behavioral vulnerabilities.