Unverified50% confidenceFactExact time
Organizations including Google DeepMind have published internal documents exploring the question of model welfare.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/16/2026
First Seen
Valid until: 12/15/2026
Sources
Related Entities
Related Claims
UnverifiedOrganizations such as OpenAI and Google DeepMind have also begun disclosing differences in model performance across different topic categories in their respective safety reports.69% similarUnverifiedResearch teams at institutions including Google and DeepMind are already exploring having models output probability distributions that reflect the perceptual tendencies of different groups.68% similarUnverifiedGoogle DeepMind在2023年提出了能力弹性（Capability Elicitation）框架，强调模型的真实行为边界只有在特定触发条件下才会暴露67% similarUnverifiedOrganizations such as Anthropic and Google DeepMind have already practiced similar 'Responsible Scaling Policies' internally.67% similarUnverifiedGoogle DeepMind的Mariner项目也在探索类似Computer Use的方向67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/925245API
curl https://kongchang.com/api/v1/knowledge/claims/925245MCP
get_claim(id=925245)