Timeline for Why do ChatGPT “jailbreaks” work?
Current License: CC BY-SA 4.0
18 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 4, 2023 at 4:20 | answer | added | Hiren Namera | timeline score: 2 | |
| Oct 27, 2023 at 17:47 | comment | added | TylerDurden | @JonathanReez yes, very well said. | |
| Oct 27, 2023 at 17:24 | comment | added | JonathanReez | Of course, note that the importance of jailbreaks has been wildly blown out of proportion by the media. If I input "death to humanity" in Word nobody accuses Microsoft of improper behavior. But if I convince ChatGPT to output "death to humanity" through some clever tricks, it's somehow turns into a huge ordeal. In a saner world OpenAI would just shrug and tell people they don't care in the slightest. | |
| Oct 26, 2023 at 9:06 | answer | added | JRE | timeline score: 7 | |
| Oct 26, 2023 at 7:33 | comment | added | Nelson | You can actually create an adversarial AI to deliberately create jailbreak prompts. One of the easiest thing to do is to try to get it to do illegal things, like suggest a top 10 piracy website. You then create an adversarial LLM with the goal of making the target LLM divulge the "forbidden" info. It is very much a cat and mouse game, but now taken to the next level. The only way for the LLM to be "safe" is have such a massive lead in processing power that any "undesirable" outputs and its respective prompts are trained away in a very short time. | |
| Oct 26, 2023 at 3:52 | comment | added | Luke Sawczak | @ArthurAttout Came here to say exactly this. | |
| Oct 26, 2023 at 1:50 | comment | added | John Gordon | Chat swear/curse filters can be defeated by using words that are close in appearance or meaning to real swear words. On a very high level, I imagine the same principles apply to jailbreaking. | |
| S Oct 25, 2023 at 19:33 | history | edited | TylerDurden | CC BY-SA 4.0 |
Tidy up post
|
| S Oct 25, 2023 at 19:33 | history | suggested | cocomac | CC BY-SA 4.0 |
Tidy up post
|
| Oct 25, 2023 at 19:24 | review | Suggested edits | |||
| S Oct 25, 2023 at 19:33 | |||||
| Oct 25, 2023 at 18:34 | vote | accept | TylerDurden | ||
| Oct 25, 2023 at 17:32 | answer | added | JimmyJames | timeline score: 25 | |
| Oct 25, 2023 at 9:26 | comment | added | Ander Biguri | In essence, a LLM is several billion numbers that are multiplied and operated with each other in a very particular way. And the numbers that the algorithm outputs, they in the end are converted into a text. There is no way to know what the numbers mean in the middle, there is no way to see the holes. You can experience them, but not see them. | |
| Oct 25, 2023 at 7:01 | comment | added | Arthur Attout | Keep in mind that AI models are mostly blackboxes. You cannot define precise reasoning steps within it. To change the behavior of an AI model, there is only one thing you can do : train it over data that depicts the behavior you want it to mimic. If despite this, there are still holes, there is next to nothing you can manually do to alter the model's output. | |
| Oct 25, 2023 at 4:32 | history | became hot network question | |||
| Oct 24, 2023 at 21:53 | answer | added | Neil Slater | timeline score: 44 | |
| S Oct 24, 2023 at 20:30 | review | First questions | |||
| Oct 25, 2023 at 4:42 | |||||
| S Oct 24, 2023 at 20:30 | history | asked | TylerDurden | CC BY-SA 4.0 |