OpenAI withdraws three math manuscripts after sign error invalidates key proof

OpenAI has withdrawn three mathematical manuscripts, including "Algebraicity of Weil classes on split abelian eightfolds," after a sign error invalidated a stabilization-trace cancellation argument, according to a Hacker News post and a history file in its openai/math GitHub repository.

The withdrawn papers are "Algebraicity of Weil classes on split abelian eightfolds," "Algebraicity of Kuga–Satake Correspondences for K3 Surfaces," and "The rational Hodge conjecture for products of K3 surfaces."

The withdrawal also affects two papers that depended on the invalidated construction. The withdrawn papers now carry notices explaining the gap and linking to archived manuscripts. The same update reportedly revised 14 other manuscripts with proof repairs and added six formalizations; the feed supplied no details on those revisions.