research into ai chatbots and prompting

my datacamp subscription expired last week but this week is free access week for ai and python so i’ll just renew it next week. today i will be looking at three courses based off of ai chatbots to help with some basckground knowledge if i decide to use llms in my final year project

here are some notes i took (i didn't take notes for everything):

understanding chatgpt, ch1: interacting with chatgpt

• limitations of chatgpt
  • trained on data form up to a certain date depending on its version
  • bias from training data
  • can struggle to track context from history of chat
• writing effective prompts
  • “garbage in,, garbage out”

understanding chatgpt, ch2: adopting chatgpt

• identifying use cases for chatgpt
  • not to require accurate results
  • the response quality can be verified
  • no input of sensitive data
  • ownership of content not required

understanding prompt engineering, ch1: prompting unveiled

• the next level
  • cot - chain-of-thought prompting
  • zero-shot - no structure provided
  • one/few-shot - roadmap of steps for chatgpt to follow to solve problems