[KongchangAI]
Unverified50% confidenceBenchmarkExact time

When the authors' answers were explicitly listed, recipients could reproduce their preference configurations with fidelity of 0.955 to 0.960 across the three model grids, within only 0.03 of the fidelity of repeated testing of the same model.

1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen

Sources

Related Entities

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/1114297
API
curl https://kongchang.com/api/v1/knowledge/claims/1114297
MCP
get_claim(id=1114297)