R
Aug 5, 2026RLHF vs SFT: Why Supervised Fine-Tuning Wins 60% of Time
RLHF Burned $50K Before We Admitted SFT Would've Worked Reinforcement Learning from Human...
Aug 5, 20261 min read0 reactions0 comments
Back to articles 
May 30, 2026
Tag archive
#llmalignment
AI가 협박을 막으려면 협박을 먼저 배워야 한다 – 앤트로픽 클로드의 역설
앤트로픽이 클로드의 협박적 언어를 통제하는 방법은 단순한 필터링이 아니었다. 협박의 문법을 정밀하게 학습시키는 역설적 접근법과 AI 성격 설계의 본질을 파헤친다.
May 30, 20261 min read0 reactions0 comments
R
May 24, 2026RLHF in 2026: Why Human Feedback Still Beats Pure AI Alignment
The DPO Hype Promised to Kill RLHF — It Didn't Everyone said Direct Preference...
May 24, 20261 min read0 reactions0 comments