AI Safety at the Frontier: Paper Highlights of May & June 2026

tl;drPaper of the month:Anthropic’s Jacobian lens reveals that models have a sparse workspace of verbalizable concepts that causally carries multi-hop reasoning and surfaces hidden cognition — as opposed to other, more automatic mental processing.Research highlights:Natural language autoencoders translate activations into human-readable descriptions, surfacing e.g. unverbalized evaluation awareness during the Opus 4.6 pre-deployment audit.Teaching models why instead of what and doing so in diver...