We believe our results are cause for cautious optimism regarding the ability to train language models to abide by ethical principles, echoing encouraging results we saw in our earlier related work on Constitutional AI (CAI).

→ View original post on X — @anthropicai