News for hackers. Stuff that matters.

A comment on Mistral Large 4. Parent comment

Re:Mistral Large 4 by imtringued on Wednesday October 07, 2026 @06:53AM (#49990957)
You got it backwards. RL is long running inference plus training at the end. They let the Agent spin for quite a while, then they calculate the reward afterwards and modify the parameters accordingly.
[ Parent | on HN ]
Re:Mistral Large 4 by fxwin on Wednesday October 07, 2026 @08:26AM (#49991805)
... so training? Even "regular" pretraining is "inference plus training at the end" if you want to frame it that way
[ Parent | on HN ]