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

Feb-12-2026, 10:19:00 GMT–Neural Information Processing Systems 

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