Report
Sword Health reportedly built mental health support model Dawn 1 on prime-rl
PrimeIntellect says Sword Health RL-post-trained GLM-5.3 into Dawn 1, claiming it is safer than its base.
TLDR
PrimeIntellect says Sword Health built Dawn 1 on prime-rl by reinforcement-learning post-training GLM-5.3. It describes the mental health support model as safer, less sycophantic and more cost-efficient than the base model.
Combined views
—
1 Source, first seen ago
