Unverified50% confidenceFactExact time
Google's research shows that emergent abilities tend to be step-change transitions; for example, GPT-3 can perform multi-step reasoning at 175 billion parameters but is almost incapable of it at 13 billion parameters.
1
Sources
50%
Confidence
Long-term
Relevance
9/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedGoogle DeepMind研究表明,涌现能力表现为非线性的相变特征,模型参数量从10B增长到100B过程中某些复杂推理能力会突然跃升70% similarUnverifiedGoogle Brain团队于2022年系统记录了涌现能力现象,发现某些能力在模型达到特定规模前几乎为零,之后突然跃升66% similarUnverifiedThe GPU resources required for a single inference of a hundred-billion-parameter flagship model may be 10 to 20 times higher than those of a ten-billion-parameter model.64% similarUnverifiedModels with larger parameter counts and more reasoning steps may consume 10–50 times the GPU compute of lightweight models.64% similarUnverifiedKimi 2.5 computes only approximately 30 billion parameters per forward pass, despite having a total parameter count of 1 trillion.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/860050API
curl https://kongchang.com/api/v1/knowledge/claims/860050MCP
get_claim(id=860050)