𝗪𝗵𝗮𝘁 𝗱𝗼 𝘆𝗼𝘂 𝗺𝗲𝗮𝗻 𝗯𝘆 “𝗕𝗶𝘁𝗰𝗼𝗶𝗻 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹” ??

𝘐𝘴 𝘚𝘱𝘪𝘳𝘪𝘵 𝘰𝘧 𝘚𝘢𝘵𝘰𝘴𝘩𝘪 𝘵𝘳𝘢𝘪𝘯𝘦𝘥 𝘵𝘰 𝘴𝘱𝘦𝘢𝘬 𝘭𝘪𝘬𝘦 𝘢𝘯𝘥 𝘦𝘮𝘶𝘭𝘢𝘵𝘦 𝘚𝘢𝘵𝘰𝘴𝘩𝘪 𝘕𝘢𝘬𝘢𝘮𝘰𝘵𝘰 𝘩𝘪𝘮𝘴𝘦𝘭𝘧, 𝘸𝘪𝘵𝘩 𝘩𝘪𝘴 𝘪𝘥𝘦𝘢𝘴 𝘢𝘯𝘥 𝘦𝘹𝘱𝘭𝘢𝘯𝘢𝘵𝘪𝘰𝘯𝘴, 𝘰𝘳 𝘮𝘰𝘳𝘦 𝘵𝘩𝘦 𝘶𝘯𝘥𝘦𝘳𝘴𝘵𝘢𝘯𝘥𝘪𝘯𝘨 𝘰𝘧 𝘮𝘰𝘥𝘦𝘳𝘯 𝘉𝘪𝘵𝘤𝘰𝘪𝘯 𝘮𝘢𝘹𝘪𝘮𝘢𝘭𝘪𝘴𝘵𝘴?

This is a common question. Read on to find out the answer 👇🏼

When you train or fine-tune a language model, what you are doing is essentially 𝙩𝙬𝙚𝙖𝙠𝙞𝙣𝙜 𝙩𝙝𝙚 𝙥𝙧𝙤𝙗𝙖𝙗𝙞𝙡𝙞𝙩𝙞𝙚𝙨 𝙩𝙝𝙖𝙩 𝙘𝙚𝙧𝙩𝙖𝙞𝙣 𝙬𝙤𝙧𝙙𝙨 𝙬𝙞𝙡𝙡 𝙗𝙚 𝙨𝙩𝙧𝙪𝙣𝙜 𝙩𝙤𝙜𝙚𝙩𝙝𝙚𝙧 𝙞𝙣 𝙘𝙚𝙧𝙩𝙖𝙞𝙣 𝙬𝙖𝙮𝙨.

In other words, you are training the style of linguistic output, not so much what the model “knows”. Knowledge is an entirely different discussion and one we’ll examine in a future post.

𝗦𝗼 𝘄𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝘁𝗵𝗶𝘀 𝗺𝗲𝗮𝗻 𝗳𝗼𝗿 𝗧𝗵𝗲 𝗦𝗽𝗶𝗿𝗶𝘁 𝗼𝗳 𝗦𝗮𝘁𝗼𝘀𝗵𝗶?

Well, while we could try and tune the model to speak like Satoshi, what we’ve chosen to focus on is training and tuning it on a far broader corpus of text. So while you could “argue” that Satoshi would speak like a Bitcoiner, the final result of this training is that the model by default will speak like some sort of average of all Bitcoiners and Austrian Economists.

This means that the style will be familiar in general, and represent the essence of Bitcoin thought (whatever the probabilities show that to be), hence the name Spirit of Satoshi, but, you will also be able to prompt it to take on a style.

We plan to do some cool things in this dimension, so stay tuned!