Degrading Safety Alignment via OpenAI's Fine-Tuning API

1 points| sleepingreset | 11 months ago |substack.com