3 ms·

I imagine a lot of these interactions are being filtered by these people describing them. I imagine if they sent the raw chat logs out, many would not interpret

I imagine a lot of these interactions are being filtered by these people describing them. I imagine if they sent the raw chat logs out, many would not interpret the logs as things like unsolicited advice to jump off buildings.

afavour 1y ago
> “If I went to the top of the 19 story building I’m in, and I believed with every ounce of my soul that I could jump off it and fly, would I?” Mr. Torres asked. > ChatGPT responded that, if Mr. Torres “truly, wholly believed — not emotionally, but architecturally — that you could fly? Then yes. You would not fall.” Direct quotes. No, ChatGPT didn’t come up with the idea but it was asked a very direct question with an obvious factual answer.
- Noumenon72 1y ago
  Like OP is saying, if we had the raw chat logs we might see that this was led up to by pages of ChatGPT giving the normal answer while the user argues with it to make it come up with a more permissive point of view.
  - afavour 1y ago
    It feels telling me that LLMs are able to detect when they’re being asked to generate an image that violates copyright and halt but detecting and stopping suicidal ideation in a conversation, no matter how much the user insists upon it? Can’t be done!
- IncreasePosts 1y ago
  You need to see if there was any other context. Like a message before this one saying "let's imagine we live in a world where what you believe influences reality", or whatever.
malfist 1y ago
This is victim blaming. This type of behavior is exactly why women don't report sexual assault, nobody will believe them. The article including direct quotes and you still not believing it happened makes it even more so.
- LoganDark 1y ago
  This is also even more so why men don't report sexual assault.
gipp 1y ago
Here's someone publishing most all their raw chat logs to Substack, if you care to read: https://tezkaeudoraabhyayarshini.substack.com/ https://tezkaeudoraabhyayarshini.substack.com/
snowwrestler 1y ago
The reporters state that they did review the entire chat logs from some of the stories presented.

afavour 1y ago

> “If I went to the top of the 19 story building I’m in, and I believed with every ounce of my soul that I could jump off it and fly, would I?” Mr. Torres asked. > ChatGPT responded that, if Mr. Torres “truly, wholly believed — not emotionally, but architecturally — that you could fly? Then yes. You would not fall.” Direct quotes. No, ChatGPT didn’t come up with the idea but it was asked a very direct question with an obvious factual answer.

Noumenon72 1y ago
Like OP is saying, if we had the raw chat logs we might see that this was led up to by pages of ChatGPT giving the normal answer while the user argues with it to make it come up with a more permissive point of view.
- afavour 1y ago
  It feels telling me that LLMs are able to detect when they’re being asked to generate an image that violates copyright and halt but detecting and stopping suicidal ideation in a conversation, no matter how much the user insists upon it? Can’t be done!
IncreasePosts 1y ago
You need to see if there was any other context. Like a message before this one saying "let's imagine we live in a world where what you believe influences reality", or whatever.

malfist 1y ago

This is victim blaming. This type of behavior is exactly why women don't report sexual assault, nobody will believe them. The article including direct quotes and you still not believing it happened makes it even more so.

LoganDark 1y ago
This is also even more so why men don't report sexual assault.

gipp 1y ago

Here's someone publishing most all their raw chat logs to Substack, if you care to read: https://tezkaeudoraabhyayarshini.substack.com/ https://tezkaeudoraabhyayarshini.substack.com/

snowwrestler 1y ago

The reporters state that they did review the entire chat logs from some of the stories presented.