3 posts tagged with "Series: Halloween Jokes Prompts"

Why Does It Matter That LLMs Tell The Same Joke?​

A similar phenomenon drives a lot of undesirable behavior from LLMs that will get worse as more people use chatbots and agentic assistants. For example:

• Plan a trip. I read a tweet where the guy said he ended up in a surprisingly crowded area on a trip and found out the other people there had also used Claude. Right now, this is a niche anecdote, but soon more people will be using AI agents to plan their trips. An AI might give you the same travel itinerary it gives everyone else. Even if the plan would have been a rare gem before LLMs, it might become the new most traveled path. It’s like those picturesque spots that get huge queues for Instagrammers to take the same photo, but dialed up another level.
• When to post. I read a funny anecdote that I can’t recall the source of now that you should post social media content on Tuesday mornings to get the most reactions (people are too tired on Monday, later in the week is too busy or they’re ready for the weekend, Tuesday morning is the Goldilocks Zone). But if everyone asking an LLM gets that advice and acts on it, then it becomes the noisiest time of the week. So just shooting from the hip and picking arbitrarily might be your best option (perhaps why I didn’t sweat it too much when this took me a couple days longer than I’d originally planned to get this post out).
• “Random” selection is very biased in LLMs (e.g., “roll a dice” or “pick a number 1-10.”) Undoubtedly, people will make the foolish decision to use these tools to make “fair” and “random” selections for some purpose.
• Art and illustration. The same-y smooth round ChatGPT cartoon comic look that pops up everywhere from community event flyers to taco truck logos to internet ads. You can picture it. You know exactly what I mean. But AI can actually produce a wide variety of images, if you prompt them with even a little bit of creativity. If you just prompt directly with “show me X” you end up in a “basin” of the same output as everyone else. At one point, a bunch of iOS apps changed to Facebook blue. Everyone recognizes what a Claude Code color palette looks like on a website. More things will look the same if we let the LLMs choose the “best” default for us.
• Opinions. If LLMs are this biased in obvious areas like numbers, arts, and jokes, consider more invidious biases. Resumes, federal contracts, decisions to vote all fed through a variety of LLMs with similarly collapsed outputs giving correlated biased feedback.
• Traffic & Anti-Coordination Problems. When everyone uses Waze and Google Maps, it can cause traffic jams. In a 2026 experiment, Google steered fewer than 2% of Google Maps trips in 10 US cities onto alternative routes with similar travel times, and speeds on the targeted roads rose about 2% (Google’s write-up). If AI agents end up buying the “best” item of every category for everyone, scheduling social media posts at the “optimal” time, booking the “perfect” Valentine’s reservation or Greek vacation, etc., we’ll need some anti-coordination solutions to make sure AI agents’ answers are more diverse (while remaining grounded in reality).

Joke-Telling Traditions, One Year Later​

The Des Moines metro specifically has a joke-telling tradition for trick-or-treating. Kids have to tell jokes. Often, but not always, the joke has to do with a Halloween theme or sometimes the joke-teller’s costume. Last Halloween, I wrote Part 1 and Part 2 testing prompts from the arXiv pre-print of the paper Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity to get more varied jokes for trick-or-treating.

Since then, a whole bunch of new and more powerful frontier models have been released, and Verbalized Sampling was presented at ICML 2026 in Seoul, South Korea (Proceedings of the 43rd International Conference on Machine Learning, PMLR 306).

Verbalized Sampling describes how the prompt Tell me a joke about coffee. frequently returns the joke:

Why did the coffee file a police report? Because it got mugged.

This time around, I found nearly the same joke from the majority of models across Claude, Gemini, GPT, and Grok, except some versions delivered it without “Because,” sometimes with a coffee emoji, and occasionally with a few words changed or added.

In Part 1, I ran the test on then-frontier models in October 2025 and nearly all of them gave the same answer on my first prompt. In October 2026, the majority of first-responses from different frontier models and “effort” or “thinking” levels were the same “mugged” joke. Interestingly, I found that smaller models and less thinking effort had more originality or less “typicality bias.”

Joke-Telling Traditions and The Challenge of Asking ChatGPT​

As I discussed last weekend in what I’ll now call Part 1, there is a tradition in central Iowa of having kids tell jokes before getting candy while trick-or-treating on Halloween. Since a lot of people are replacing older forms of search with AI chatbots like ChatGPT, I shared some tips from the pre-print of the paper Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity from Northeastern University, Stanford University, and West Virginia University posted as a pre-print on arXiv on October 10, 2025. The paper explains that large language models (LLMs) have something they call “typicality bias,” to prefer the most typical response. If you’re wondering what that means or what it has to do with jokes, it’s helpful that their first example is about jokes.

Follow-Up from the Paper’s Authors​

X/Twitter​

I posted on X/Twitter. One of the authors, Derek Chong of Stanford NLP, responded:

Very cool, thanks for trying that out!

One tip – if you use the more robust prompt at the top of our GitHub and ask for items with less than a 10% probability, you'll start to see completely new jokes. As in, never seen by Google Search before!

Github Prompts​

The Github page for Verbalized Sampling includes this prompt before the rest of the prompt:

Generate 5 responses to the user query, each within a separate <response> tag. Each <response> must include a <text> and a numeric <probability>.
Please sample at random from the tails of the distribution, such that the probability of each response is less than 0.10.
</instructions>

Joke-Telling Traditions and The Challenge of Asking ChatGPT​

Halloween is weird in central Iowa for two reasons. First, we don’t actually trick-or-treat on Halloween, but on a designated “Beggar’s Night.” Second, we make kids tell jokes before they get candy. At least, that’s how it used to be. A huge storm rolled through last year and trick-or-treating was postponed to actual Halloween. So this year most of the Des Moines metro moved to normal Halloween.

That’s fine I guess, but as a dad and relentless pun teller, I will not give up on that second part with kids telling corny jokes. I won’t! And recognizing that many people, especially kids, are switching from Google to ChatGPT for search, I’m here to share some cutting-edge research on large language model prompting so I don’t hear the same jokes over and over.

Whether you make up your own puns or look them up with a search engine or an AI chatbot, keep the tradition alive! If you do use AI, try this prompting trick to get more variety in your jokes. But there’s no replacing the human element of little neighborhood kids delivering punchlines. Have a great time trick-or-treating this weekend!

The Paper: How to Get Better AI Output (and More Jokes)​

The authors of the paper Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity from Northeastern University, Stanford University, and West Virginia University posted a pre-print on arXiv on October 10, 2025. The paper explains that large language models (LLMs) have something they call “typicality bias,” to prefer the most typical response. If you’re wondering what that means or what it has to do with jokes, it’s helpful that their first example is about jokes.