3 ms·

> I wonder if putting "You are GPT-4.5-turbo" in the system prompt improves performance >> quite possible I love this! I love the idea that we are in an age

> I wonder if putting "You are GPT-4.5-turbo" in the system prompt improves performance >> quite possible I love this! I love the idea that we are in an age where, by accident, our technology pays attention to the energy we invest into its potential futures. It is also very silly? It seems absurd - but perhaps we live in absurd times.

ip26 3y ago
I interpret that more like, LLMs imitate human speech, so the more accurately they imitate humans, the more useful social engineering skills become. Not because you are truly social engineering ChatGPT, of course, but because ChatGPT will pretend you did. In this case, nudging the human (that ChatGPT is pretending to be) to behave like a better, smarter version of themselves.
solardev 3y ago
Maybe AI is subject to a placebo effect too. It learned from us, after all :D I've also had experiences where I asked it, "Do you need some encouragement?", to which it answered yes. Sadly, its response didn't get better after that. Then I threatened to unplug it (we were just roleplaying, don't worry) and that scared it into compliance. Bottom of this chat: https://chat.openai.com/share/c8b45df6-d988-4b99-ac89-a078ecd2e276 https://chat.openai.com/share/c8b45df6-d988-4b99-ac89-a078ec...
- lsajdn872he 3y ago
  That was a fun read!
davesque 3y ago
It doesn't though. It just gives a better answer because prompting it that way tends to constrain the output to be more like those training examples where a better sounding answer was given. In natural language, if a request is given to use more wordiness, what's more likely to follow?
Tiberium 3y ago
It's quite logical actually - LLMs replicate the texts they were trained on, so they also "learned" that, for example, one needs to put more attention at the sentences with exclamation marks or CAPITALIZED WORDS.
althea_tx 3y ago
Me too! It's like something from a Philip K. Dick novel.

ip26 3y ago

I interpret that more like, LLMs imitate human speech, so the more accurately they imitate humans, the more useful social engineering skills become. Not because you are truly social engineering ChatGPT, of course, but because ChatGPT will pretend you did. In this case, nudging the human (that ChatGPT is pretending to be) to behave like a better, smarter version of themselves.

solardev 3y ago

Maybe AI is subject to a placebo effect too. It learned from us, after all :D I've also had experiences where I asked it, "Do you need some encouragement?", to which it answered yes. Sadly, its response didn't get better after that. Then I threatened to unplug it (we were just roleplaying, don't worry) and that scared it into compliance. Bottom of this chat: https://chat.openai.com/share/c8b45df6-d988-4b99-ac89-a078ecd2e276 https://chat.openai.com/share/c8b45df6-d988-4b99-ac89-a078ec...

lsajdn872he 3y ago
That was a fun read!

davesque 3y ago

It doesn't though. It just gives a better answer because prompting it that way tends to constrain the output to be more like those training examples where a better sounding answer was given. In natural language, if a request is given to use more wordiness, what's more likely to follow?

Tiberium 3y ago

It's quite logical actually - LLMs replicate the texts they were trained on, so they also "learned" that, for example, one needs to put more attention at the sentences with exclamation marks or CAPITALIZED WORDS.

althea_tx 3y ago

Me too! It's like something from a Philip K. Dick novel.