Has anyone considered (or already completed) testing models to see how well they can answer questions about this filtered data?
Consider an eval that tests how well models' awareness of the alignment literature. "Models know everything about misaligned AI, including benefits and strategies for scheming" seems like a useful thing for us to know? Or is it a foregone conclusion that they'd saturate this benchmark anyway?

Really appreciate you pulling this together Ruby. Amazing work.
Two resources I've found useful or interesting that weren't on the above list:

1. Emotion-focused therapy has huge effect sizes for increasing martial satisfaction (suspiciously huge). Even if you account for publication/other biases I'd still wager it's one of the best rehabilitate and preventative interventions for couples (but it basically just pools lots of what you've discussed, like turning toward hard conversations and non-violent communication). This is the best book summarising it: https://www.amazon.com.au/Hold-Me-Tight-Conversations-Lifetime/dp/1491513810
2. My PhD student did a systematic review of randomised experiments to help people make relationship decisions (e.g., stay or go? if that link is still being minted, see here). Basically, there's nothing, except for helping decide whether to leave an abusive partner. But, the one study on non-abusive couples was interesting: if you flip a coin and do what the coin says, you're much more likely to be happy six months later. The most obvious explanation here is status quo bias: people are afraid of leaving but are happier once they do. If you're thinking hard about leaving, such that you'd be willing to do what a coin says, then probably just go. If you're not quite ready, I found the idea of setting tripwires (see Ch.11 summary here) helpful to make it clear what I want to see change without constantly shifting the goalposts on myself (e.g., if x doesn't happen in 3 months, I'm out).

Obviously a different market and more mainstream but the Alliance for Decision Education has funding, big names on their board, and plans in motion to scale rationality training (under a different name) https://alliancefordecisioneducation.org/

Thanks jimv. Fixed