News for hackers. Stuff that matters.
Re:GPT-5.2 by XCSme on Friday December 12, 2025 @05:37AM (#46242794)
But most benchmarks are not about that...

Are there even any "hallucination" public benchmarks?

Re:GPT-5.2 by andrepd on Friday December 12, 2025 @06:13AM (#46243002)
"Benchmarks" for LLMs are a total hoax, since you can train them on the benchmarks themselves.
Re:GPT-5.2 by XCSme on Friday December 12, 2025 @02:40PM (#46247893)
I would assume a good benchmark has hidden tests, or something randomly generated that is harder to game