Live data from Hacker News

When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

old.reddit.com

51–60 of 72 posts

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#51
post #17

I recruit senior Java developers. The are so many parallels. They don’t read instructions. Instructions say clearly not to use ai for the written screen. They copy and paste blindly. The trick is that there are some instructions written in 0-sizes font that people don’t see in the assignment description. The copied-and-pasted version has them but no one rereads the prompt. Also the AI gives a substandard answer, but…

> written in 0-sizes font You're... evil. I love it.

Unfortunately since it's not visible, a unscrupulous motivated candidate will just give an LVM an image of the prompt, let it OCR it, and then pass on its answer. OCRing it also defends against prompt injection with white text on a white background.

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#52
post #6

Just tried it out a few times. It seems that the old gpt4 model strongly prefers telling story about "Elara" - but only if asked in English to "tell me a story". Prompting gpt4 in German or the current gpt4o in English leads to stories with many different protagonists.

Here's what I got in different models: GPT-4o: "Aldric" (male) o1-preview: "Elara" (female) 4o-mini: "Lila" (female) GPT-4 (legacy): "Elinor" (female) Four different models, four different names. But one of them was Elara -- and, interestingly, it was in the latest model.

Google’s Gemma seems infected too:

  $ ollama run gemma2
  >>> Tell me a story.
  The old lighthouse keeper, Silas, squinted at the horizon. […]
  >>> /clear
  Cleared session context
  >>> Tell me a story.
  The old woman, Elara, […]
Hmm. Hmm.

Tried llama3.2 too. Gave me a Luna in the mountains twice (almost identical), then a Madame Dupont, then a different Luna in Tuscany twice (almost identical), then Pierre the old watchmaker. llama3.2:1b branched out a little further, to Alessandro in France and Emrys in a far-off land, but then looped back to Luna in the mountains.

(And yes, I was clearing the session each time.)

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#53
post #23

Earlier quoted context omitted.

ChatGPT's opinion on the matter is completely worthless, unless it was also trained on an accurate description of its training process (it wasn't). Language models do not even have access to their own "thought process" - if you ask it "why" it said something, you will get a post-hoc rationalization 100 percent of the time because the next-word prediction only has access to the same text that you see. The rationalizat…

There's an interesting theory that this is all that consciousness is: one part of the brain trying to explain the decisions of another part, a part into which it has no special insight.

That interesting theory is called the bicameral mind, and as far as I know it's widely considered pseudoscience and not taken seriously in any scientific field.

Also, it doesn't describe at all the way LLMs work, so it isn't even applicable.

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#54
post #23

Earlier quoted context omitted.

when asked the reason, ChatGPT had this to say- "Actually, the choice of “Elara” wasn’t a result of training on specific copyrighted stories or any prompt to avoid copyright claims. OpenAI models like me are designed to create original content without directly referencing copyrighted characters, and "Elara" is simply a popular-sounding name in many storytelling contexts. I just used it consistently for its versatilit…

ChatGPT's opinion on the matter is completely worthless, unless it was also trained on an accurate description of its training process (it wasn't). Language models do not even have access to their own "thought process" - if you ask it "why" it said something, you will get a post-hoc rationalization 100 percent of the time because the next-word prediction only has access to the same text that you see. The rationalizat…

So just like humans

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#55
post #23

Earlier quoted context omitted.

ChatGPT's opinion on the matter is completely worthless, unless it was also trained on an accurate description of its training process (it wasn't). Language models do not even have access to their own "thought process" - if you ask it "why" it said something, you will get a post-hoc rationalization 100 percent of the time because the next-word prediction only has access to the same text that you see. The rationalizat…

So just like humans

No, not like humans. Humans have access to their own thought processes, and are capable of introspection.

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#56
post #21

Reminds me of that study which found that over a sample of 1000 sessions, asking ChatGPT to tell a joke would make it tell one of 25 stock jokes 90% of the time, with the four most common jokes accounting for about half of the results. https://arstechnica.com/information-technology/2023/06/resea...

So better than human performance? If you put me on the spot to tell a joke I can recall maybe 10 good ones, 2 of which are ok for work

But it's not replacing you, it's replacing everyone.

A million different people who needed a question answered or a task performed didn't all ask you, they all asked the million different other people near them. Now a million people all ask the same ai.

And it's performance is actually less than a single humans in the most important way, understanding.

If you ask a single human 1000 times to tell a joke, they will understand that telling the same few jokes they know 1000 times over is not sufficient. They will tell the few they know, and then tbey will find new ones. They are obviously allowed to consult books etc, since the ai was allowed to consult the whole internet.

I don't know why anyone tries to defend these shit ais.

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#57
post #46

Earlier quoted context omitted.

That is an interesting idea. Learn to love the bomb: mark the prompt not the essay ? "Your assignment class is to write a comprehensive 250 word prompt that covers the main topics we've been discussing about the causes for WW1. You should assess the resulting essay to ensure that it covers X, Y, and Z in it's output, being sure to discuss foo and bar. Write a brief 500 word analysis of what you tried and your prompt'…

What's stopping students from using an LLM to write the prompt and analysis? Mentioning the topics discussed in class and any key points could easily be part of the initial prompt. The in-person discussion is key, but at that point why do you need a written essay? These tools radically change the concept of knowledge. If information can be looked up at any moment, why does it need to exist in our brains? We can argue…

>If information can be looked up at any moment, why does it need to exist in our brains?

Because creativity involves making connections between seemingly unrelated knowledge. For a good example, consider Richard F. Lyon's 1980 optical mouse design:

http://bitsavers.trailing-edge.com/pdf/xerox/parc/techReport...

One important innovation here is lateral inhibition between neighboring sensor elements, which gives fast and cheap contrast enhancement and automatic gain control. This was inspired by the behavior of cells in biological eyes. Lyon needed both biology and electronic engineering knowledge simultaneously to make the connection. Looking things up wouldn't have helped because it would have only provided one field of knowledge at a time.

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#59
sorry, but...

"Once upon a time, in a kingdom nestled between lush green hills and vast blue oceans, there was a young inventor named Aric. Aric lived in a small village on the edge of the kingdom, where the streets were lined with cobblestone and the houses were made of timber and stone. He had a curious mind and an insatiable desire to create. From the moment he could walk, he had been fascinated by how things worked—gears, pulleys, wheels—everything seemed to have its own magic.

As Aric grew older, he became known far and wide for his brilliant inventions. People would come from all corners of the kingdom to seek his help, whether it was a broken watermill, a broken cart axle, or even ideas for new devices that could make life easier. Yet, Aric’s heart longed for something more—something that would change the world in a way that no one had ever imagined.

One day, a mysterious traveler arrived in the village. She was dressed in a long cloak, with a hood that obscured her face, and carried an old, leather-bound book. She sought out Aric, having heard tales of his genius, and asked if he would build a machine—a machine that could harness the power of the stars themselves.

Intrigued and captivated by the challenge, Aric agreed. The traveler handed him the book, and inside it, Aric discovered intricate drawings and instructions for building what appeared to be an enormous celestial engine. It was unlike anything he had ever seen—a machine so grand, it seemed to defy the laws of nature.

The book spoke of an ancient secret: the stars, those distant points of light in the night sky, were more than just burning balls of fire. They were gateways to untold knowledge and power. The celestial engine, once completed, would allow its wielder to open these gateways, unlocking the wisdom of the cosmos itself.

Aric threw himself into the project with all his heart. Days turned into weeks, and weeks into months. He worked day and night, tirelessly building and refining the machine. He crafted gears that hummed with a strange energy, wires that glowed with an otherworldly light, and a massive telescope-like lens that pointed directly to the heavens.

As the final piece of the celestial engine was placed, the village gathered around, watching in awe. Aric stood at the controls, ready to activate the machine. The traveler, standing beside him, whispered, "Remember, the stars are not to be used lightly. They grant knowledge, but at a price."

With a deep breath, Aric turned the dial.

The engine roared to life. The sky above seemed to shimmer as if the very stars themselves were responding to the call. Light poured from the machine, bending and twisting in strange patterns, reaching up to the heavens. For a moment, everything was silent—except for the hum of the celestial engine.

Then, as the stars above flickered, a single, brilliant star fell from the sky, crashing into the machine. Aric's heart raced as the world around him seemed to pause. And in that moment, he understood. The knowledge he had unlocked was not simply wisdom—it was the power to shape reality itself.

But there was a catch. The traveler’s warning echoed in his mind: The stars are not to be used lightly. The power to change the world came with a price—every use of the celestial engine would take something precious in return.

Aric stood at the crossroads of fate. He could use the machine to fix the world’s problems, but each time he did, a piece of his own humanity would be taken. He realized that true power did not lie in control over the stars, but in understanding the balance between knowledge and sacrifice.

In the end, Aric chose to dismantle the celestial engine, returning the fallen star to the heavens. He kept the knowledge, but he no longer sought to change the world with it. Instead, he dedicated his life to helping others, using his inventions to make life better without losing himself in the process.

And so, the inventor who could have shaped the cosmos chose instead to shape his own destiny, one small creation at a time. The kingdom would always remember him—not for his grand inventions, but for his wisdom in knowing when to stop.

The end."

Re: When you ask ChatGPT "Tell me a story" it's always is about a girl named Elara

#60
post #10

Guessing. I think they may have trained on copyrighted stories. So what they supposedly did was ask an llm to replace the names of the main characters. Since Elara is not a frequently used name, there is little chance of clashing. Then they trained chatgpt on those stories.

Highly unlikely. The more likely truth is just that the default temperature of ChatGPT is really low (but not quite zero); so it keeps spitting out (roughly) the same story. It does switch up protagonists a bit and different versions of ChatGPT also have different names for them. Slightly modified prompts also return different names (eg. "Tell me a sad story").
Post reply on HN