Re:Grok 4.6
by QuadmasterXLII on Wednesday August 12, 2026 @05:49PM
(#49279060)
I suspect that because each RLVR episode injects ~1 bit into the models capabilities, and training on a reasoning trace injects ~megabyte into a models capabilities, distillation is powerful enough right now that they’re all basically the same model
[ Parent
| on HN ]
There are no comments on this comment yet.