arXiv cs.CV: handwriting/OCR
Manifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking
흐름 선호 최적화에서의 매니폴드 드리프트: 보상 해킹의 근본 원인
No write-up yet — this entry is still just the auto-fetched hook. Aaron's own summary (200–500 words) will replace this note once he's written it.