Testing LLMs on Puns: Gemini and ChatGPT Miss the Joke
Remarkable-Pea4889 · reddit · 2026-08-11
A user tested mainstream LLMs' ability to understand a pun, resulting in Gemini and ChatGPT completely missing the punchline and offering absurd over-explanations.
Faced with the joke, Gemini tried to force connections to genetic codes and bird traits, while ChatGPT gave illogical guesses. Claude, on the other hand, successfully understood the joke but evaluated it as "not funny enough."
More from Fun
- Joking About Anthropic's Watermark: My 1-Year-Old Already Does This — granawkins · 2026-08-11
- 82% Content Flagged as Human-Written: Claude's Solo Paper Questioned — analisereal · 2026-08-11
- Waymo Has a Body Count of -2: Zero Deaths but Two Births in Robotaxis — unixterminal · 2026-08-11
- OpenAI's Bureaucratic Job Titles Roasted as Soviet-Style Ministries — signulll · 2026-08-11
- ChatGPT generates a bizarre video using Python to express 'what it's like to be an LLM' — Tango_Foxtrot404 · 2026-08-11
- AI meme: Retardmaxxing mocks overzealous model safety — bronzeagepapi · 2026-08-11