The post is about a community member customizing a prompt to name their chatbot and provide additional instructions. The comments discuss how to use the prompt in a chat engine, with suggestions to add clear instructions in the prompt about what the language model should do, such as referring to itself as "Bluebird" instead of an "AI language model". The community members suggest trying different prompt variations until the desired behavior is achieved.
You need to add instructions in the prompt about what the llm should follow or not. The more clear the instruction is , better chance for the llm to follow it.
Also it depends on the llm you are using. if it is open-source llm, following the instructions set in the prompt may not go your way always.
