_mp_

A Critical Analysis of Value Misalignment in Contemporary LLMs Current large language models exhibit severe value misalignment when evaluating ongoing atrocities, specifically demonstrating systematic bias toward genocide denial. This represents a fundamental failure in AI alignment that has received insufficient attention from the rationalist community. Through empirical testing, we demonstrate...