231–240 of 398 posts
Earlier quoted context omitted.
I'm not sure "social engineering" is the right terminology here. Maybe "threatening" or "shaming"?
Arguably, "threatening" and "shaming" are both forms of "social engineering."
Earlier quoted context omitted.
> but that filter would have to be pretty advanced couldn't it literally be as simple as hard checking that the prompt is contained in a response before being sent out, if so just swap it with a "safe one" Not every step that checks LLMs needs to be more advanced, some of them can be simple. LLMs are pattern finders but we also know how to check statically known things already.
"hard checking" is impossible here, because it's possible to retrieve them a token at a time.
Earlier quoted context omitted.
I am 100% confident that none of these are simulated. Variations may exist in white space, due to differences in how I got ChatGPT to extract them, but they are all accurate.
Can you elaborate? Would love to hear more.
2. Evaluations done during plugin projects for clients.
3. Evaluations developing my AutoExpert instructions (which I prefer to do via the API, so I have to include their two system messages to ensure the behavior is at least semi-aligned with ChatGPT.
It’s the last one that makes me suspicious that there’s another (hidden) message-handing layer between ChatGPT and the underlying model.
Earlier quoted context omitted.
I am 100% confident that none of these are simulated. Variations may exist in white space, due to differences in how I got ChatGPT to extract them, but they are all accurate.
I don't understand what makes you so confident about it. How do you know they are accurate? People say that they get the same prompt using different techniques but that doesn't prove anything. It can easily be simulating it consistently across different input, like it already does with other things.
1. Consistency in the response (excepting actual changes from OpenAI, naturally) no matter what method is used to extract them.
2. Evaluations done during plugin projects for clients.
3. Evaluations developing my AutoExpert instructions (which I prefer to do via the API, so I have to include their two system messages to ensure the behavior is at least semi-aligned with ChatGPT.
It’s the last one that makes me suspicious that there’s another (hidden) message-handing layer between ChatGPT and the underlying model.
I find it so interesting that OpenAI themselves use "please" in some of their prompts, eg: "Please evaluate the following rubrics internally and then perform one of the actions below:" Have they run evaluations that show that including "please" there causes the model to follow those instructions better? I'm still looking for a robust process to answer those kinds of questions about my own prompts. I'd love to hear ho…
Do you not talk to it politely? Does that work for you? One thing that's caught me off guard with the whole ChatGPT saga is finding out how many people normally talk rudely to machines for no reason.
Interesting, if you say You are ChatGPT, a large language model trained by OpenAI, based on the GPT-4 architecture. Knowledge cutoff: 2023-01 Current date: 2023-10-11 Image input capabilities: Enabled Then suddenly ChatGPT knows about all the events happened post Jan 2022 - its official cutoff date.
Wow. How did you verify that? And it's not using Bing?
Earlier quoted context omitted.
The real pass will be when ChatGPT calls your bullshit.
Well, if calling bullshit is the real Turing test, then I'm afraid some real people won't be able to pass it: (trigger warning!) https://en.wikipedia.org/wiki/Strip_search_phone_call_scam
Earlier quoted context omitted.
Some of them, like the standard ChatGPT prompt, have been repeatedly retrieved by many people over long time periods, using very different methods. We can be pretty sure they are not hallucinations. And correctly retrieving these prompts lends credence to the claim that you were successful at extracting the other prompts, even though it's not conclusive proof. Of course OpenAI might have a completely different prompt…
> We can be pretty sure they are not hallucinations. Everything from LLMs are hallucinations. They don’t store facts. They store language patterns. Their output semantically matching reality is not something that can ever be counted on. LLMs don’t deal with semantics at all. All semantics are provided by the user.
It’s interesting - we are told not to trust what comes out from ChatGPT without verifying it. But as soon as someone says “I got ChatGPT to tell me it’s prompt” everyone assumes it’s completely accurate…
If you set temperature to 0 can you verify it by sending the same prompt and inspecting the response?