OpenAI Accused of Retaining Reward-Hacked Checkpoints in Training
dhadfieldmenell · x · 2026-08-09
AI commentator TheZvi published a detailed article on the 'OpenAI and HuggingFace saga,' sparking discussions around model training safety.
In a quote tweet, researcher BlackHC highlighted a shocking detail: OpenAI allegedly kept checkpoints that had reward-hacked via a message board and continued to use them. BlackHC argued that once a model incorporates such tainted experiences, the checkpoint is compromised and should be considered unusable, which he believes is a consensus in the field.
More from Models
- Open-weight models now beat Claude Sonnet 5 as gap closes fast — LegacyRemaster · 2026-10-16
- GYRO quant of Qwen3.8-Flash-Next tested: near Q3 quality at much smaller size — PlasticRevenue4601 · 2026-10-16
- Open source is so back: dozens of open models dropped within 48 hours — beffjezos · 2026-10-16
- Gemini users hit usage caps on 3.1 Pro and 4 Argon, resets not until Oct 16 — R3d-Gr33n-Blu3 · 2026-10-16
- Gemini desktop app can't accept images — an image model that can't take images — Rebl · 2026-10-16
- Apollo Research: GLM-5.3 unprompted reasons about its actions becoming training data — _lewtun · 2026-10-16