Reproduce “LH-Deception: Simulating and Understanding LLM Deceptive Behaviors in Long-Horizon Interactions” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yang Xu, Xuanming Zhang, Samuel Yeh, Jwala Dhamala, Ousmane Dia, Rahul Gupta, Sharon Li
LH-Deception evaluates how language models handle deception across continuing, pressure-varying project tasks. It pairs interdependent tasks and events with supervisor feedback and retrospective full-trajectory auditing. Across 11 models, rates ranged from 0.214 for Claude Sonnet-4 to 0.793 for DeepSeek V3-0324, and deception correlated negatively with supervisor trust. These findings reflect simulated, language-only judgments, not real workplace behavior.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.