Pulling the data that exist in the context, obviously. The name is baked in during training, and this training usually comes last and is always extremely shallow, so LLMs often get lost there. But if your LLM fails simple context retrieval, it's probably already too confused for any other task.
Pulling the data that exist in the context, obviously. The name is baked in during training, and this training usually comes last and is always extremely shallow, so LLMs often get lost there. But if your LLM fails simple context retrieval, it's probably already too confused for any other task.