TASM Notes 011

Fri Mar 15, 2024

Bit late on this one; it's actually the meeting notes for last week.

Pre-Meeting Chatting

Zvi's Update

The Talk - Mesaopetimizers and Robustness

Term Check

Interest: "Robustness Guarantee"

The Paper: Risks from Learned Optimization

Goal Misgeneralization

Concrete example:


What's the difference between Robustness and Alignment?

Mesaoptimizers/inner misalignment in the wild

Pub Time

Presumably, they all discussed the pubbed items from above, but I didn't end up joining this time.

Creative Commons License

all articles at langnostic are licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License

Reprint, rehost and distribute freely (even for profit), but attribute the work and allow your readers the same freedoms. Here's a license widget you can use.

The menu background image is Jewel Wash, taken from Dan Zen's flickr stream and released under a CC-BY license