Not all LLMs are born to be chatbots

What is instruction fine-tuning

There is a common misconception that large language models like GPT-3 are chatbots that can naturally hold conversations.

However, LLMs are not inherently conversational — they are trained to generate fluent text, not to answer questions or take instructions. LLMs need additional fine-tuning to handle tasks like answering questions and dialogue.

For exam…