DeepSeekR1, the LLM that everyone seems to be talking about has a similar flaw especially when asked any prickly questions ...