Talk:OpenAI o3

Latest comment: 6 days ago by Alenoach in topic GPQA Diamond

Feedback from New Page Review process

edit

I left the following feedback for the creator/future reviewers while reviewing this article: Great start! I've added a few more sources to ensure a WP:GNG pass which requires multiple independent articles. :)

MolecularPilot 🧪️✈️ 02:53, 22 December 2024 (UTC)Reply

GPQA Diamond

edit

Thanks for the article but no explanation what GPQA - not even to talk about GPQA Diamond benchmark means. Could be anything for non-AI people. 79.142.230.127 (talk) 17:46, 27 December 2024 (UTC)Reply

Thanks for the feedback, but I don't know if we can be more precise in the article without digressing too much. And GPQA doesn't seem notable enough for a separate article, as far as I can tell. Perhaps we could indicate what the abbreviation GPQA means (Graduate-Level Google-Proof Q&A), if it doesn't make the sentence too cluttered.
To explain it here, GPQA Diamond is GPQA's "highest quality subset which includes only questions where both experts answer correctly and the majority of non-experts answer incorrectly"[1] Alenoach (talk) 06:00, 28 December 2024 (UTC)Reply