Unverified50% confidenceFactExact time
OpenAI has acknowledged that GPT-6 Astra's monitorability has declined, attributing this solely to the model becoming smarter rather than to architectural changes or deliberate training
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/4/2026
First Seen
Valid until: 1/2/2027
Sources
Related Entities
Related Claims
UnverifiedWhen handling highly sensitive tasks such as training frontier large models, Opus 5.5 will automatically fall back to a weaker version.72% similarUnverifiedOpenAI在InstructGPT论文中证明，少量高质量人类偏好数据对模型对齐的贡献远超大规模无监督训练72% similarUnverifiedThe safety alignment in the original training of open-source models can be removed through a small amount of fine-tuning.70% similarUnverifiedOpenAI omitted key parameters such as model scale and training compute from the GPT-4 technical report, citing 'competitive considerations and safety factors.'69% similarUnverifiedTools such as the Training API lower the barrier to custom training, enabling teams with high-quality proprietary data to convert their data advantage into a model advantage.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/972619API
curl https://kongchang.com/api/v1/knowledge/claims/972619MCP
get_claim(id=972619)