> Those examples, if included in training data, would confuse an LLM and likely lead to poor results. I don't think that LLMs are good enough that they can the

> Those examples, if included in training data, would confuse an LLM and likely lead to poor results. I don't think that LLMs are good enough that they can they confused by logical inconsistencies in the training data.