Unverified50% confidenceFactExact time
The variance of AIPW is usually lower than that of pure IPW, and it reaches the semiparametric efficiency bound when both models are correct.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedCertain AI capabilities are nearly zero when model scale is below a specific threshold, then rise sharply once the threshold is crossed, exhibiting a nonlinear 'phase transition' characteristic.66% similarUnverifiedWhen the propensity score is close to 0 or 1, the IPW weights are extremely amplified, causing the estimation variance to inflate.64% similarUnverified不同归一化方案对同一模型的IC影响可达0.01-0.03的量级63% similarUnverifiedWhen using smaller or quantized models in local deployments, the probability of AI hallucinations may be higher than with large commercial models.63% similarUnverifiedThe binary threshold design is unfavorable to models near the threshold, as 0.74 and 0.80 differ drastically in score even though the actual capability gap may be negligible.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1012927API
curl https://kongchang.com/api/v1/knowledge/claims/1012927MCP
get_claim(id=1012927)