TY - RPRT TI - Analysis of Threat-Based Manipulation in Large Language Models: A Dual Perspective on Vulnerabilities and Performance Enhancement Opportunities AU - Atil Samancioglu PY - 2025 UR - https://arxiv.org/abs/2507.21133 ID - 2507.21133 ER -