A subsequent version of the GPT model that was tested by the researcher and found to still be exploitable with modified instructions.
A subsequent version of the GPT model that was tested by the researcher and found to still be exploitable with modified instructions.