[flagged]
Show HN: Unicode Steganography
11–18 of 18 posts
Re: Show HN: Unicode Steganography
#12Cool stuff. I think there have been projects recently that use LLMs to encode messages in plain text by manipulating the choices of output tokens. Someone with the same version of the LLM can decode. Note sure where to find these projects though.
Wow, thaHt’s soELP interestiIMng… weA wouLIld lovVEe toTR heaAPPr morEDe aboINut thaUSEAST1t topic!
(With apologies to Mr Justice P. Smith, sort of: https://en.wikipedia.org/wiki/Smithy_code )
Re: Show HN: Unicode Steganography
#13You can actually do better: hint - variational selectors, low bytes.
Switching between NFC vs NFD could be even sneakier.
Re: Show HN: Unicode Steganography
#14Super interesting. Any examples in the wild of any state of the art models responding to this?
Re: Show HN: Unicode Steganography
#15Cool stuff. I think there have been projects recently that use LLMs to encode messages in plain text by manipulating the choices of output tokens. Someone with the same version of the LLM can decode. Note sure where to find these projects though.
This is a really interesting space, and one that I've been playing with since the first GPTs landed. But it's even cooler than simply using completion choice to encode data. It has been mathematically proven that you can use LLMs to do stego that cannot be detected[0]. I'm more than positive that comments on social media are being used to build stego dead drops. What I find really interesting about this approach is t…
Very neat
Re: Show HN: Unicode Steganography
#16[flagged]
Re: Show HN: Unicode Steganography
#17[dead]
Re: Show HN: Unicode Steganography
#18Cool stuff. I think there have been projects recently that use LLMs to encode messages in plain text by manipulating the choices of output tokens. Someone with the same version of the LLM can decode. Note sure where to find these projects though.
How long until this gets to the point that we can play Doom over LLM conversations?