Large language models (LLMs) sometimes learn the wrong lessons, according to an MIT study. Rather than answering a query based on domain knowledge, an LLM could respond by leveraging grammatical patterns it learned during training. This can cause a model to fail unexpectedly when deployed on new tasks.Large language models (LLMs) sometimes learn the wrong lessons, according to an MIT study. Rather than answering a query based on domain knowledge, an LLM could respond by leveraging grammatical patterns it learned during training. This can cause a model to fail unexpectedly when deployed on new tasks.[#item_full_content]