42. Seems like I managed to independently confirm Aurora results on SYNTH (600M parameters). Very early run but promi… by @Dorialexander (Alexander Doria) · backlist 2026-05-09 · rubric 88.0
40. i guess that's probably where prime is heading, but i could see synth with RLM as a general RL alternative for lo… by @Dorialexander (Alexander Doria) · backlist 2026-05-08 · rubric 92.0
64. Ok directly relevant for ongoing work (on memorization): avoiding a "huge percentage of neurons to effectively di… by @Dorialexander (Alexander Doria) · backlist 2026-05-08 · rubric 88.0