petergpt/bullshit-benchmark
BullshitBench measures whether AI models challenge nonsensical prompts instead of confidently answering them, created by Peter Gostev.
3 best points · 0 comments · 3 Hacker News posts · 1,906 GitHub stars · View on GitHub
Hacker News discussions
- Bullshit Benchmark: how do chatbots respond to silly questions? (3 points, 0 comments, by twistorial, 2026-02-25)
- BullshitBench v2: LLMs answering nonsense questions (2 points, 0 comments, by nkko, 2026-03-05)
- BullshitBench: Detect nonsense, call it out and avoid confidently continuing (2 points, 0 comments, by fittingopposite, 2026-03-20)