Following last year's in-depth article by @OpenAI's former safety systems team lead @lilianweng on LLM hallucinations, OpenAI's latest paper once again discusses LM Hallucinate (language model hallucinations)