Is code-davinci-002 just the largest non-GPT-4 model in the GPT-4 scaling law experiment? | Manifold

Is code-davinci-002 just the largest non-GPT-4 model in the GPT-4 scaling law experiment?

Basic

3

Ṁ35

Feb 17

44%

chance

1D

1W

1M

ALL

https://arxiv.org/pdf/2303.08774.pdf

Resolves Yes/No when I have >90% confidence.

This question is managed and resolved by Manifold.

Get

1,000

and

3.00

Sort by:

It's 4 orders of magnitude less compute than GPT-4! GPT-3 was 2 OOMs less compute.

Related questions

What will be true about GPT-5?

What will be true about GPT-5? (See description)

Will the performance jump from GPT4->GPT5 be less than the one from GPT3->GPT4?

Will GPT-5 have more than 5 trillion parameters?

What hardware will GPT-5 be trained on?

Will GPT-4 be trained (roughly) compute-optimally using the best-known scaling laws at the time?

Will GPT-4 improve on the Chinchilla scaling law?

Is GPT-4 best? (Thru 2025)

Will GPT-5 have more than 10 trillion parameters?

What is the main reason behind GPT-4o speed improvement relative to GPT-4 base model?

Related questions

What will be true about GPT-5?

Will GPT-4 be trained (roughly) compute-optimally using the best-known scaling laws at the time?

What will be true about GPT-5? (See description)

Will GPT-4 improve on the Chinchilla scaling law?

Will the performance jump from GPT4->GPT5 be less than the one from GPT3->GPT4?

Is GPT-4 best? (Thru 2025)

Will GPT-5 have more than 5 trillion parameters?

Will GPT-5 have more than 10 trillion parameters?

What hardware will GPT-5 be trained on?

What is the main reason behind GPT-4o speed improvement relative to GPT-4 base model?

© Manifold Markets, Inc.•Terms + Mana-only Terms•Privacy•Rules