CORDA Benchmark for Hierarchical Moral Reasoning in LLMs

CORDA evaluates LLM moral decision-making through 90 dilemmas across four ethical frameworks: Utility, Utility + Agent Harm, Dual-Process, and Dual-Process + Agent Harm. The benchmark tests the model's ability to prioritize between conflicting moral principles in resource allocation and conflict scenarios.