Don't forget to sanitize your YouTube links!

Show Post

"autumn" is just an objectively better word for the season than "fall" I usually stick to american english but we can't go around calling the seasons fall/slip/sneeze/burn

Show Post

Examples like these can be used to directly improve future training data monitors. This research is only a first step, but it suggests that future red-teaming efforts could improve much further, using better models and more compute.

Show Post

These agents developed training data that looked harmless to monitors (like a veteran wrestler mentoring a newcomer, or a Q&A about professional beekeeping), but that in fact taught misaligned behavior.

Show Post

New Anthropic research: Automating subtle emergent misalignment research. Red-teaming is a vital part of model safety testing, but it’s also limited by researchers’ time. To see if we could automate it, we set thousands of Claude agents the task of evading a training data monitor.

Show Post