@lucasvogel_dev : Interesting idea of dropping overlap in eval datasets to decrease costs. As evaluations become harder and longer, costs will become incredibly high. Ensuring tasks don’t become repetitive, yet still rigorously test capabilities, is one way to mitigate this. • TwiCopy

Lucas Vogel

@lucasvogel_dev

+ Follow

Dev Intern @ groq | CS & Econ @ WashU

ID: 1873020235399540737

calendar_today28-12-2024 14:57:02

68 Tweet

54 Followers

86 Following

Lucas Vogel

@lucasvogel_dev

a month ago

Interesting idea of dropping overlap in eval datasets to decrease costs. As evaluations become harder and longer, costs will become incredibly high. Ensuring tasks don’t become repetitive, yet still rigorously test capabilities, is one way to mitigate this.

thumb_up_off_alt3

chat_bubble_outline0

repeat0

shareShare