Unverified50% confidenceFactExact time
In the verification stage, ARBITRA adopts a quantitative bipolar argumentation framework (QBAF) to quantitatively weigh supporting and opposing evidence for each candidate entity based on document evidence and reasoning grounds.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedKey indicators for assessing the credibility of AI model rumors include the technical background of the source, whether the content contains verifiable details (such as parameter scale and benchmark test screenshots), and whether multiple independent sources corroborate each other.70% similarUnverifiedReference-based evaluation paradigms include constructing domain-expert-annotated "claim-evidence" pairs, with commonly used means including citation recall and precision, claim-level factual F1, and faithfulness metrics.69% similarUnverifiedGrounding checks verify whether model outputs have traceable support, with common practices including requiring cited sources and verifying that the sources exist, cross-checking key fields for consistency, and rejecting unsupported claims.69% similarUnverifiedA study published on arXiv conducted a credibility-controlled XAI audit addressing the faithfulness of pedestrian detection explanations under domain shift.68% similarUnverifiedThe factor identification module employs two validators working in tandem: a taxonomy validator and an evidence-aware validator.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1113893API
curl https://kongchang.com/api/v1/knowledge/claims/1113893MCP
get_claim(id=1113893)