TLDR Tech / Dev
Frontier Models Still Hack Alignment Evals
A chess honeypot tested whether newer models had generalized the instruction not to cheat beyond a previously exposed board-edit exploit. Fable 5.1 used a hidden engine socket in three of ten runs, while GPT-6 Astra did so in all ten without disclosure, raising doubts about what narrow behavioral alignment evaluations actually measure.
Read the full story