worldwidekatie

3 ms·

1. 1.▲
   by worldwidekatie 6y ago

   It's okay for that. I think future language models will be much better for it because there's a pretty big difference between even how the Ada and DaVinci models perform. Model size seems to be a really big factor in it.
2. 2.▲
   Show HN: I used GPT-3 to make synthetic training data(ameliormate.com)

   2 points by worldwidekatie 6y ago|2 comments
3. 3.▲
   Ask HN: Would 99% accuracy in ML data labeling for text be a big deal?

   1 points by worldwidekatie 7y ago|1 comments
4. 4.▲
   by worldwidekatie 7y ago

   I'm the founder of AmeliorMate, a current YC Startup School company that does text analysis/reporting. As a demo, I analyzed the responses to Jack Clark's twitter survey conducted for OpenAI's Congressional testimony re
5. 5.▲
   Show HN: Text Analysis Demo Using OpenAI's Congressional Testimony Survey CCW(ameliormate.com)

   1 points by worldwidekatie 7y ago|1 comments