News for hackers. Stuff that matters.

A comment on Even GPT-5.2 Can't Count to Five: Zero-Error Horizons in Trustworthy LLMs. Parent comment

Re:Even GPT-5.2 Can't Count to Five: Zero-Error Horizons in ... by azakai on Thursday April 02, 2026 @02:29PM (#47618306)
You are trying it on a production model. The paper is using models with tool calls disabled.
[ Parent | on HN ]

There are no comments on this comment yet.