Will a “good” (fine-tuned) opensource model utilizing chain-of-thought reasoning like o1 be released by EOY 2024?
➕
Plus
20
Ṁ1600
Dec 31
37%
chance

The model should be able to decide how long it needs to think based on the complexity of the problem. Ideally it should be ranked higher on LMSYS than the “normal” model but this is not a strict requirement. I imagine OLAMA would update their UI to support something like that but also would be ok if someone else does it. If model isn’t ranked higher on LMSYS then the market would resolve based on my experience playing around with it. In general, it has to be significantly better than just using the normal version. Ameaturish attempts will be dismissed. I won’t be betting on this market.

Get
Ṁ1,000
and
S3.00
© Manifold Markets, Inc.Terms + Mana-only TermsPrivacyRules