You are off to a whirlwind start with this blog! You think carefully about problem that evidently has a jackpot of leverage to solve, run a very revealing study from which you extract plenty of great data from 98 days of meals, and make a steel case for why LLM companies should take this whole situation very seriously. The argument that animal-heavy diet recommendations are a misalignment with Claude’s constitution ties all the other great arguments together wonderfully and with force, and even if some other company were only interested in preventing gender stereotypes, they have plenty of reason now to put your findings to work, as well. This is a lot of excellent effort gathering evidence and rationale that I imagine can, and hopefully will, make fixing these staggeringly influential models plenty easier. I just very much hope they see your great work and proposition! 🐔️
You are off to a whirlwind start with this blog! You think carefully about problem that evidently has a jackpot of leverage to solve, run a very revealing study from which you extract plenty of great data from 98 days of meals, and make a steel case for why LLM companies should take this whole situation very seriously. The argument that animal-heavy diet recommendations are a misalignment with Claude’s constitution ties all the other great arguments together wonderfully and with force, and even if some other company were only interested in preventing gender stereotypes, they have plenty of reason now to put your findings to work, as well. This is a lot of excellent effort gathering evidence and rationale that I imagine can, and hopefully will, make fixing these staggeringly influential models plenty easier. I just very much hope they see your great work and proposition! 🐔️
(Also the link just now stopped working, sadly)