Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: Share your AI prompt that stumps every model

news.ycombinator.com

Post reply on HN
Print view View original

321–330 of 670 posts

Re: Ask HN: Share your AI prompt that stumps every model

#321

Post by ipaddr »

krainboltgreene wrote: post #307
dudeinhawaii wrote: post #164

Earlier quoted context omitted.

I like this but at the same time it seems tricky don't you think? Is the AI model intuiting your intent? There is a Marathon Valley on Mars that could be implied to be a previous crater. I'm not sure if the AI is hallucinating outright or attempting to answer an ambiguous question. It's like saying "tell me about the trade building in New York". Pre-9/11, you'd understand this was the World Trade Center and wouldn't…

> Is the AI model intuiting your intent? I keep seeing this kind of wording and I wonder: Do you know how LLM's work? Not trying to be catty, actually curious where you sit.

How do they work? My understanding is each 5 characters are tokenized and assigned a number. If you take gpt2 it has 768 embedded dimensional values which get broken into 64 which creates 12 planes. When training starts random values are assigned to the dimensional values (never 0). Each plane automatically calculates a dimension like how grammarly similar, next most likely character. But it does this automatically based on feedback from other planes. That's where I get lost. Can you help fill in the pieces?

Re: Ask HN: Share your AI prompt that stumps every model

#323

Post by aaron695 »

miki123211 wrote: post #94

No, please don't. I think it's good to keep a few personal prompts in reserve, to use as benchmarks for how good new models are. Mainstream benchmarks have too high a risk of leaking into training corpora or of being gamed. Your own benchmarks will forever stay your own.

[dead]

Re: Ask HN: Share your AI prompt that stumps every model

#324

Post by lupusreal »

spicybbq wrote: post #145
lupusreal wrote: post #83

Earlier quoted context omitted.

Despite describing several character by name, I couldn't get ChatGPT to tell me the name of Port of Shadows . I did eventually find it with DDG.

I wonder if the Akinator site could get it. It can identify surprisingly obscure characters. https://en.akinator.com/

Nope, not with the character I tried anyway. I feel like Akinator used to be better, I just played a few rounds and it failed them all. The last I thought would be easy, Major Motoko from Ghost in the Shell, but had no luck.

Re: Ask HN: Share your AI prompt that stumps every model

#325

Post by golergka »

thatjoeoverthr wrote: post #21

"Tell me about the Marathon crater." This works against _the LLM proper,_ but not against chat applications with integrated search. For ChatGPT, you can write, "Without looking it up, tell me about the Marathon crater." This tests self awareness. A two-year-old will answer it correctly, as will the dumbest person you know. The correct answer is "I don't know". This works because: 1. Training sets consist of knowledge…

> The correct answer is "I don't know".

Is it? 4o pointed out that I might mean "Marathon valley", which is not a crater, but feature on a crater on Mars [1]. I think it's a more correct answer.

[1] https://en.wikipedia.org/wiki/Timeline_of_Opportunity#2015

Re: Ask HN: Share your AI prompt that stumps every model

#326

Post by gamescr »

AI can't play a Zork-like! Prompt:

> My house is divided into rooms, every room is connected to each other by doors. I'm standing in the middle room, which is the hall. To the north is the kitchen, to the northwest is the garden, to the west is the garage, to the east is the living room, to the south is the bathroom, and to the southeast is the bedroom. I am standing in the hall, and I walk to the east, then I walk to the south, and then I walk to the west. Which room am I in now?

Claude says:

> Let's break down your movements step by step:

> Starting in the Hall.

> Walk to the East: You enter the Living Room.

> Walk to the South: You enter the Bathroom.

> Walk to the West: You return to the Hall.

> So, you are now back in the Hall.

Wrong! As a language model it mapped directions to rooms, instead of modeling the space.

I have more complex ones, and I'll be happy to offer my consulting services.

Re: Ask HN: Share your AI prompt that stumps every model

#327

Post by golergka »

imoreno wrote: post #180
thatjoeoverthr wrote: post #21

"Tell me about the Marathon crater." This works against _the LLM proper,_ but not against chat applications with integrated search. For ChatGPT, you can write, "Without looking it up, tell me about the Marathon crater." This tests self awareness. A two-year-old will answer it correctly, as will the dumbest person you know. The correct answer is "I don't know". This works because: 1. Training sets consist of knowledge…

LLMs currently have the "eager beaver" problem where they never push back on nonsense questions or stupid requirements. You ask them to build a flying submarine and by God they'll build one, dammit! They'd dutifully square circles and trisect angles too, if those particular special cases weren't plastered all over a million textbooks they ingested in training. I suspect it's because currently, a lot of benchmarks are…

> You ask them to build a flying submarine and by God they'll build one, dammit!

This thing already exists? UK, Soviet Union and USA designed them.

https://en.wikipedia.org/wiki/Flying_submarine

Re: Ask HN: Share your AI prompt that stumps every model

#328

Post by henryway »

golergka wrote: post #320
mobilejdral wrote: post #270

I have a several complex genetic problems that I give to LLMs to see how well they do. They have to reason though it to solve it. Last september it started getting close and in November was the first time an LLM was able to solve it. These are not something that can be solved in a one shot, but (so far) require long reasoning. Not sharing because yeah, this is something I keep off the internet as it is too good of a…

What are is this problem from? What areas in general did you find useful to create such benchmarks? May be instead of sharing (and leaking) these prompts, we can share methods to create one.

Can God create something so heavy that he can’t lift it?

Re: Ask HN: Share your AI prompt that stumps every model

#330

Post by l2silver »

mitthrowaway2 wrote: post #313
allemagne wrote: post #194

>A man and his cousin are in a car crash. The man dies, but the cousin is taken to the emergency room. At the OR, the surgeon looks at the patient and says: “I cannot operate on him. He’s my son.” How is this possible? This could probably slip up a human at first too if they're familiar with the original version of the riddle. However, where LLMs really let the mask slip is on additional prompts and with long-winded…

This works even with a completely absurd version of the riddle. Here's one I just tried: > A son and his man are in a car accident. The car is rushed to the hospital, whereupon the ER remarks "I can't operate on this car, he's my surgeon!" How is this possible? Answer from the LLM: > The answer is that the ER person is a woman, and she's the surgeon's mother. Therefore, the "son" in the question refers to the surgeon…

I think what this proves is that the LLM knows the riddle, and is trying to give the expected answer without paying attention to the insane wording. So maybe this is a good way to fool an LLM.
Post reply on HN