Hacker News Reader

except they have similar pretrain/rlhf data which is the thing u really want to tune for

YMMV but for me even models in the same family fail in different ways, and every incremental update changes it