The more accurate your Prompts are, the closer the responses will be to what you need. OpenAI offers six steps, and notes that the methods can be combined for ‘greater impact.’
#1 Write Clear Instructions
Want short responses? Ask for them. Want less simplistic responses, request that. Using specific requests in your instructions will allow a system like ChatGPT to know exactly what you want. After all, these chatbots are not really mind readers.
OpenAI Tips for Improving Clarity of Instructions:
Include as many details as possible in the instruction to receive more relevant responses
Ask the model to function as a specific persona, for example – behave like a data analyst and create a report for me….
Use separators to specify separate parts of the input
Outline the steps required to complete a task
Provide examples
Specify the desired length of the output
#2 Provide Auxiliary Text
Systems like OpenAI’s ChatGPT are good – but that doesn’t mean they are perfect. Even the strongest models sometimes return less accurate responses than expected. To improve this, OpenAI suggests instructing the model to respond based on the auxiliary text you provide.
#3 Break Down Complex Tasks into Simpler Subtasks
Instead of uploading a document to ChatGPT and asking it to do everything at once, the OpenAI guide explains that better results can be achieved by breaking down the desired tasks.
For example, instead of uploading a report and asking the chatbot to summarize the report, ask it to specify specific things in several tasks: summarize the chapter on expenses, summarize the revenues, etc.
#4 Give the Model Time to ‘Think’
Patience is always good. OpenAI suggests including in the instruction a request for the model to show its ‘chain of thought,’ allowing the model to take more time and sharpen the response while thinking. You can also ask the model to review the task to ensure it doesn’t miss anything.
#5 Use External Tools
You can integrate ChatGPT with external tools to improve the results. ChatGPT has a vast number of add-ons, such as OpenTable for finding great restaurants or Kayak for easily planning your next vacation.
The OpenAI guide explains that add-ons or external tools can ‘compensate for the model’s weaknesses.’
‘If a task can be performed more reliably or efficiently by a tool rather than a language model, use it to get the best out of both,’ the OpenAI guide states.
For the full guide including examples and tactics click here