I'm also working on a library to steer the sampling step of LLM's but more for steganographic / arbitrary data encoding purposes.
Should work with any llama.cpp compatible model: https://github.com/sutt/innocuous
Should work with any llama.cpp compatible model: https://github.com/sutt/innocuous
Watch how the "Cumulative encoding" row grows each iteration (that's where the BTC address will be encoded) and then look at the other rows for how the algorithm arrives at that.
Thanks for checking it out!