Unveiling the Bias Impact on Symmetric Moral Consistency of Large Language Models

Nov-17-2025, 01:21:59 GMT–Neural Information Processing Systems 

We find that the ability of LLMs to maintain consistency varies across different moral scenarios.