Alex Mallen

mentions 2 type Person feed RSS

// recent coverage 2 mentions

22:54

2026-07-31

lesswrong.com

ai-safety

SOTA alignment assessments don’t strongly update us against misalignment

Anthropic's April 2025 alignment risk update for its Claude Mythos Preview model concluded that the model 'does not possess any unknown propensities that would increase alignment risk,' but a LessWron…

21:21

2026-06-18

lesswrong.com

ai-safety

The distillation double bind: Distilling misaligned models either transfers misalignment or it doesn't

Researchers propose distillation techniques to transfer AI capabilities without transferring misalignment, leveraging a double bind where failed incrimination distillation reduces misalignment risk. T…

// co-occurs with top 8 entities

Anders Woodruff 2 Buck Shlegeris 2 Carlo Leonardo Attubato 1 Eric Gan 1 Aniket Chakravorty 1 Francis Rhys Ward 1 Julian Stastny 1 Anthropic 1

// topics top 5 topics

ai safety 2 ai research 2 machine learning 1 large language models 1 ai policy 1