Unverified50% confidenceFactExact time
In a single call, Jev simultaneously answers four questions: whether it may be phishing or a scam, whether the visible link text points to a different website, whether it has spam or misleading characteristics, and an overall judgment of whether it is safe/suspicious/scam, with each answer accompanied by a probability value.
1
Sources
50%
Confidence
Long-term
Relevance
10/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn a prompt injection resistance test, Jev maintained the technical support classification by treating the message content as untrusted text, keeping the refund probability at 3%.63% similarUnverifiedJev is suitable for high-value classification scenarios such as enterprise invoice fraud detection, spam filtering, community content moderation, high-frequency refund request handling, and customer churn prediction.60% similarUnverifiedMultiple questions in Jev are evaluated in parallel, so sending many questions takes roughly the same amount of time as sending a single question.58% similarUnverifiedThe debugging order for wrong RAG answers is: 1) is the correct information in the index; 2) was it retrieved; 3) is its ranking high enough to enter the prompt; 4) did the model actually use it. Most bugs appear in the first three steps.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1049418API
curl https://kongchang.com/api/v1/knowledge/claims/1049418MCP
get_claim(id=1049418)