3 ms·
The comparison between base pretrained models and the RLHF production/helpful models suggests that it's likely reinforcement learning which is developing the in
by benlivengood 1y ago
The comparison between base pretrained models and the RLHF production/helpful models suggests that it's likely reinforcement learning which is developing the introspection.