One would expect on priors that they trained the models in similar ways, including for alignment. They have the same training cutoff date. If so, the divergence suggests the base mode or other kinds…

Posters: repligate. repligate

2 likes recorded by the source corpus.