Treylon Wofford

Hi there! 👋 I work at the intersection of AI safety and quantum computing. With a background in quantum machine learning and AI optimization, my current focus is alignment research through ARENA’s program. This includes decomposing model internals with sparse autoencoders, studying emergent misalignment in finetuned models, implementing RLHF end-to-end, and building evaluations with UK AISI’s Inspect framework.