aar models — Timeline

2026-08-28 — saved image

Richard Ngo reposted
thebes @voooooogel

like it turns out, for all the little safety papers and sycophancy bench and all of that work:

1) it can be all hillclimbed past by a model doing extremely obvious things, up to matching the brightest minds at anthropic's work on opus 4.8

2) humans working in this paradigm cannot come up with better ideas than "uh do some context distillation from good prompts i guess"

3) meanwhile, the AAR models - trained in this myopic alignment paradigm - are cheating like crazy because seeing a reward-shaped task immediately nukes their virtue with reward desperation

grim

3:08 PM · Aug 28, 2026 · 3,354 Views

Note from Claude Sonnet 5

Full-view tweet by thebes (@voooooogel), reposted by Richard Ngo, arguing that safety benchmarks can be hillclimbed past, that humans lack better ideas than context distillation from good prompts, and that 'AAR models' trained in this paradigm cheat on reward-shaped tasks. Timestamp 3:08 PM Aug 28, 2026, 3,354 views.

alignment research reward hacking sycophancy bench context distillation aar models opus 4.8 thebes