Unverified50% confidenceFactExact time
The system compares each hop of each returned path against the source sentences, labeling them as one of three categories: 'source-supported', 'inferred assumption', or 'insufficient evidence'.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEnd-to-end prompting methods entangle the three components of evidence retrieval, reasoning, and uncertainty estimation, making it difficult to pinpoint the source of errors.68% similarUnverifiedProvenance belongs to the 'source verification' line of defense, does not rely on patterns for identifying injected text, and is therefore also robust against unknown variants.67% similarUnverifiedIn multi-step tool-calling chains, errors can propagate and amplify between reasoning nodes, making root-cause analysis difficult.66% similarUnverifiedThe system prompt requires all responses to cite sources, and when retrieval results are irrelevant, it must respond with 'I don't know.'66% similarUnverifiedWhen a query requires multiple pieces of evidence, any-hit records 'only getting part of the evidence' and 'getting the evidence in full' as the same result, leading to inflated metrics and distorted comparison between approaches.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1063247API
curl https://kongchang.com/api/v1/knowledge/claims/1063247MCP
get_claim(id=1063247)