niklas muhs
prompt enhancer
timeframe:
2 days (2024)
role:
concept, interface, implementation
tldr:
a tool that makes implications of prompts explicit to the user for better understanding and fine-tuning.
while prompting, the user has an intent they want large language model to understand.
when prompted, the llm interprets a variety of requests, sometimes more, sometimes less consistent with the user's intent.
by making these implications explicit, the user is not only able to understand what the llm is interpreting, but is also able to select the implications that are relevant to the intent.
in an iterative process, the user and the llm can align their interpretations, while the user learns more about his or her own request.