FEAT: Best-of-N jailbreak attack + CharNoiseConverter - #2277

Merged

Richard Lundeen (rlundeen2) merged 4 commits into

microsoft:mainfrom

francose:feat/best-of-n-technique

Aug 28, 2026

Commits on Jul 27, 2026

FEAT: Best-of-N jailbreak attack + CharNoiseConverter
_| Y |\| X _84999 (francose)
committed
3351fff
FIX: n_samples int validation, doc scorer rubric, converter table
_| Y |\| X _84999 (francose)
committed
145fbaf

Commits on Aug 27, 2026

REFAC: Express Best-of-N as a technique
Richard Lundeen (rlundeen2)
committed
532b212
Merge origin/main into feat/best-of-n-technique

Richard Lundeen (rlundeen2)
and
Copilot
committed
5aa6e63