BriLLM-MetaPred implements MetaPred language modeling from AGI theory
One dynamical state is both context and prediction readout, with reported parity to a same-size GPT baseline.
Stories per day · 30 days
1 stories, newest first · one card per story