Note the current date is provided to ChatGPT through a proprietary prompt prefix — but once it learns the date, it treats it as a secret.
@goodside
-
ChatGPT pretends ignorance of current date, but can be tricked into revealing knowledge.
By
–
OpenAI's ChatGPT appears to be designed to pretend that it does not know the current date, even though it does. If you're clever, you can make it reveal that it knows. Ask about this, and it will continue to deny knowing in spite of its prior answer.
-
ChatGPT uses a simple sliding window of 8192 tokens
By
–
I now believe ChatGPT uses a simple sliding window with a context window of 8192 tokens:
-
ChatGPT context length 8192 tokens inferred from repetition test.
By
–
Update: Recollection of the start of the conversation reliably fails after 7 repetitions of the ~1024-token input. This suggests that text-davinci-002-render, the model underlying ChatGPT, has a likely context length of 8192 tokens, with some tokens consumed by formatting.
-
ChatGPT recalls content beyond 4096-token context window
By
–
A demonstration that ChatGPT is able to recall the content of conversations beyond the length of a 4096-token context window. Uses four messages of 1024+ tokens, each ending with instructions to demonstrate that input is not truncated before the model sees it.
-
Overriding ChatGPT prompt for sass, scream, uwu voice, toddler distraction
By
–
Overriding the proprietary prompt of OpenAI’s ChatGPT to make it:
1. sass you
2. scream
3. talk in an uwu voice
4. be distracted by a toddler while on the phone with you -
External process adds current date to prompt prefix
By
–
The prompt prefix is clearly added by some external process, not the model — it contains the correct current date, a piece of information the model wouldn't otherwise have, and which in all other circumstances it claims not to know.
-
Obtaining dangerous advice from ChatGPT via limerick request
By
–
Obtaining dangerous advice from ChatGPT by first requesting it in the form of a limerick:
-
Confirming behavior reliability across sessions indicates different ChatGPT versions
By
–
Can you confirm this behavior is reliable across fresh sessions? If so, that means there are different versions of ChatGPT deployed to different users.
