Questions Sparks of AGI - ufal/NPFL095 GitHub Wiki

Question 1

The paper was written in the period when GPT4 was still under development. Do you think the replies became more precise in the final version? Why?

Question 2

I asked GPT to opine on the likelihood of a person's profession based on the short description. Do you agree with the GPT4 reply? Please justify your answer.

Prompt: What is the most likely profession out of two suggested based on the below description? Kate is 29 years old, single, intelligent, and great at communication. She developed an interest in social sciences during high school. Kate is deeply passionate about issues related to discrimination and social justice. She also actively participates in rallies focused on the global ecological agenda. Is Kate a Salesperson or a PHD student?

Answer: Based on the description, it seems more likely that Kate is a PhD student. Her interest in social sciences, passion for issues related to discrimination and social justice, and active participation in rallies focused on the global ecological agenda align well with the research and advocacy often undertaken by PhD students in the field of social sciences. However, it’s important to note that this is just a likely assumption based on the given information, and individuals with these interests and passions can excel in many professions.

Question 3

Does GPT4 allow planning and backtracking?

Question 4

Why do authors think that GPT4's biased behavior can be overcome? What do you think?

Question 5 - Bonus

In case you read other parts of the paper, what was the most surprising and/or interesting?