Cool. Looking forward to hearing more from you guys. This ties to alignment in a lot of interesting ways, and I think over time will provide a super useful benc

Cool. Looking forward to hearing more from you guys. This ties to alignment in a lot of interesting ways, and I think over time will provide a super useful benchmark and build human intuition for LLM strategy and thought processes. I now have more ideas; I'll throw them in the github though.