humans have much to learn from robotkind
it's all about the prelude before the conversation. You need to tell it what the AI is and is not capable. It's not trying to be right, it's trying to complete what it thinks the AI would do :)
I wonder if the AI would be better at math if you told it to show it's work
Teaching GPT-3 to do a brute force 'for loop' checking answers also seems to work

Jul 17, 2020 · 12:12 PM UTC

Some rejected scenarios proposed by GPT-3 while trying to avoid the chore of math.
1- A Lovecraftian black mass of ominous voices surround and scream at Holo that she is bad at math.
Unbelievable. What was your starting prompt?? In my attempts it tends to narrate the result of the dialogue instead of the answer to the question (e.g. "You ask 'what is a good approximation of pi?'" --> "She gives you an answer that's good enough to use on a test."
I tried to recreate it again to check how well it replicates. General context prompt ends in the line 'You say "So, f(x)=x*x".'. You might need to tweak the temperature to make the math more consistent.
In 2020 people developed an AI which needs a sense of purpose and positive emotional background to solve math.
The AI yelling “first try” on his second try is definite proof that it's becoming human.
Unrelated but, this is ridiculously cute the way it communicates. Would die to protect it
Hold on a second. Is this all real? I am not nearly as impressed by the function composition or arithmetic as by the metacognition and question asking. So the story-Socratic format is how to get interactivity? Really? How many trials is this? Did you really not edit this?
This tweet is unavailable