Unverified50% confidenceFactExact time
Every output from Jev comes with a calibrated probability; when Jev gives a probability of 0.8, it means it is correct approximately 80% of the time on that type of judgment.
1
Sources
50%
Confidence
Long-term
Relevance
10/6/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the fourth round of fusion experiments, treating JEV as the 9th framework and taking the highest score yielded an overall accuracy of 81.5%, but parity dropped back to 50%.71% similarUnverifiedJev can support confidence-based tiered decision-making: automatically execute above 90%, transfer to human review between 60% and 90%, and do not execute below 60%.71% similarUnverifiedThe hosted model Jev is significantly more accurate on 9 out of 11 decision points, with margins ranging from +10.8 to +46.0 percentage points.71% similarVerifiedIn a well-calibrated model, an output score of 0.8 should statistically correspond to an actual accuracy of 80%.68% similarUnverifiedJev captured 81.7% of false positive samples in terms of false alarm detection.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/980242API
curl https://kongchang.com/api/v1/knowledge/claims/980242MCP
get_claim(id=980242)