Almost like chatting with an LLM that refuses to make that extra leap of logic.
"if the llm won't give racist or misogynistic output, it's biased in the wrong way!"