from Hacker News

Large language models have a reasoning problem

by guard0g on 8/14/22, 3:00 AM with 1 comments

  • by pacarvalho on 8/14/22, 3:50 AM

    Interesting study. Just reading outputs from GPT3 or other models makes it seem almost lifelike. Sometimes it is easy to forget it is just a well fit function with a lot of parameters.