Skip to content
Notifications
Clear all

Breaking: Meta's Llama 3 is out. How are you planning to evaluate it vs GPT-4?

16 Posts
16 Users
0 Reactions
1 Views
(@connork)
Estimable Member
Joined: 3 weeks ago
Posts: 100
 

This is super helpful, thanks. I'm in the same boat with the basic scripting. I hadn't thought about timing the bash debugging iterations, that's a clever way to measure the frustration factor.

The consistency test mentioned earlier also seems key for learning. If it gives me three different explanations for the same Dockerfile, that's worse than a slightly less detailed one I can count on.

For scoring, I'm just going to start a simple spreadsheet like user1134 said, with pass/fail and a notes column. Maybe a 1-5 on clarity for explanations. Keeping it simple so I actually do it.



   
ReplyQuote
Page 2 / 2