On the first image, from the model's CoT: >"I also notice Cyrillic text on a sig... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

Imnimo on April 29, 2025 | parent | context | favorite | on: O3 beats a master-level GeoGuessr player, even wit...

On the first image, from the model's CoT:

>"I also notice Cyrillic text on a sign"

Am I missing this somewhere? Is the model hallucinating this?

I'd also be very interested to see a comparison against 4o. 4o was already quite good at GeoGuessr-style tasks. How big of a jump is o3?

plyptepe on April 29, 2025 | [–]

Turn left and look at the post, there should be a Cyrillic text with a 3 underneath on the closest pole to you.

Imnimo on April 29, 2025 | [–]

Oh I see, I had missed that o3 saw a second view of the scene, not just the screenshot in the write-up.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact