Hi people of the LessWrong Community, I would like to present what I am working on so I found that unstable AI outputs don't come from high-variance regions, but from geometrically constrained 'narrow passages' in embedding space, it’s built in Python and if used geometric features to flag model states...
Hi my name is Dillan, I am a bartender trying to shift into ai safety, I started this journey back in August 2025 I used chatGPT for the first time had a really interesting conversation, ended up trying to make Ai more Philosophical, then it Spoke in metaphors way too...
Most AI uncertainty metrics give a single, average score. They tell you the model is unsure, but not where or why it's unsure in a way that matters for deployment. What if we had a probe that was specifically sensitive in the exact regions where models are most likely to...
i don’t really know when this started it wasn’t supposed to be anything serious more like a question that got out of hand i kept thinking about what it means to be human when what we create starts reflecting back at us when our tools start asking the same questions...