# Why Do Frontier LLMs Tell Bad Jokes?

**URL:** <https://discuss.huggingface.co/t/why-do-frontier-llms-tell-bad-jokes/176995>\
**Category:** Research\
**Created:** [June 20, 2026, 10:51am UTC](https://discuss.huggingface.co/t/why-do-frontier-llms-tell-bad-jokes/176995 "2026-06-20T10:51:51Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Young25](https://avatars.discourse-cdn.com/v4/letter/y/b38774/32.png) [@Young25](https://discuss.huggingface.co/u/Young25)\
**Post date:** [June 20, 2026, 10:51am UTC](https://discuss.huggingface.co/t/why-do-frontier-llms-tell-bad-jokes/176995/1 "2026-06-20T10:51:51Z")

</div>

I built a leaderboard where humans vote for the funniest answer, and LLMs try to predict the same choice.

Models are ranked by how often they agree with human humor preferences.

> **[LLM Humor Arena: Human-Preference Ranking - a Hugging Face Space by Young25](https://huggingface.co/spaces/Young25/llm-humor-ranking-leaderboard)**
>
> This app runs a React‑based web interface directly in your browser. Just open the page and the JavaScript loads the interactive UI for you. No extra input is needed; you’ll see the full web experie...

Do you think this leaderboard can help identify which models are better suited for applications where user enjoyment, engagement, and entertainment matter?