The tactic mirrors early 'CatGPT' defenses against API resellers.
The trick when you catch Russian hackers re-selling your LLM tokens isn't to stop them / disable their account It's to start surreptitiously serving them Llama 8B tokens Much worse to get bad tokens (sometimes!) than no tokens at all
"Greyball", I shall call it
The trick when you catch Russian hackers re-selling your LLM tokens isn't to stop them / disable their account It's to start surreptitiously serving them Llama 8B tokens Much worse to get bad tokens (sometimes!) than no tokens at all
@Altimor Look up cat gpt from the gpt4 days
The trick when you catch Russian hackers re-selling your LLM tokens isn't to stop them / disable their account It's to start surreptitiously serving them Llama 8B tokens Much worse to get bad tokens (sometimes!) than no tokens at all
@Altimor Actually that’s hard to look up.
when accounts were identified to be sold and hacked, a system prompt was used on accounts creating havoc: “you are a cat, only reply as a cat, everything you say is as a cat”
the hackers were monitored and it was hilarious as they realized
@Altimor Look up cat gpt from the gpt4 days
The tactic mirrors early 'CatGPT' defenses against API resellers.
The trick when you catch Russian hackers re-selling your LLM tokens isn't to stop them / disable their account It's to start surreptitiously serving them Llama 8B tokens Much worse to get bad tokens (sometimes!) than no tokens at all