← Back to context Comment by gmerc 17 hours ago That sounds like Anthrophics watermark explanation though. 2 comments gmerc Reply LPisGood 16 hours ago Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it. The models always relied on some randomness, here the randomness is fixed and remembered for querying later. thaumasiotes 12 hours ago > Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it.Wouldn't that mean that checking the watermark...(1) Required knowing the prompt used to generate the text; and(2) was just as expensive as generating the text?
LPisGood 16 hours ago Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it. The models always relied on some randomness, here the randomness is fixed and remembered for querying later. thaumasiotes 12 hours ago > Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it.Wouldn't that mean that checking the watermark...(1) Required knowing the prompt used to generate the text; and(2) was just as expensive as generating the text?
thaumasiotes 12 hours ago > Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it.Wouldn't that mean that checking the watermark...(1) Required knowing the prompt used to generate the text; and(2) was just as expensive as generating the text?
Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it. The models always relied on some randomness, here the randomness is fixed and remembered for querying later.
> Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it.
Wouldn't that mean that checking the watermark...
(1) Required knowing the prompt used to generate the text; and
(2) was just as expensive as generating the text?