omg I cant even believe what happened with Claude ๐คฏ its like how can an AI model be so messed up ๐ seriously though, these examples are super concerning they make you question if we're ready for this kind of tech yet? and yeah, more research is def needed to create better interpretability tools...