fx2-cmix (current Hutter prize leader) wasn't tested on enwik8, but is 4% smaller than starlit, which compressed enwik8 to 15,215,107 bytes. The best result is 14.9 MB by NNCP using a transformer network on a GPU. http://mattmahoney.net/dc/text.html
-- Matt Mahoney, [email protected] On Sat, Oct 18, 2025, 5:35 PM <[email protected]> wrote: > Good news: I added my energy priming mechanism back just to see, and I > just scored for 100,000 bytes in 23,815 losslessly compressed. It brought > it down from 24,230. That's 415, and I can probably get more out of it. > > If we naively look at my naive chart of what this may result in for all of > it fed in: > 24,000 > 17,758,051 > 23,000 > 17,018,132 > > So possibly ~17,550,000 for 100,000,000 input bytes. > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + > delivery options <https://agi.topicbox.com/groups/agi/subscription> > Permalink > <https://agi.topicbox.com/groups/agi/T6cf3be509c7cd2f2-M400154be9017c9a653a72567> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T6cf3be509c7cd2f2-Mcef7ad030179813d510d9e26 Delivery options: https://agi.topicbox.com/groups/agi/subscription
