Unverified50% confidenceFactExact time
About four-fifths of the unsupported claims corresponded to numbers present in the agent's own files, but from the wrong experiment, the wrong metric, or the wrong run batch.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedManual auditing found that about one third of the sampled relationships were bound to the wrong entity target.78% similarUnverifiedAmong the 35 genuine failure cases, only 3 did not involve any assertions, indicating that the vast majority of failures stem from accepting biased statements rather than from missing information.75% similarUnverifiedA new analysis finds that most numerical errors in AI-written papers are misattributions—citing values from the wrong experiment, wrong metric, or wrong run in one's own experimental records—rather than deliberate fabrication.74% similarUnverifiedIn a standalone test of the "view incomplete" prompt on 100 questions, the wrong-answer rate was 16% without the prompt and dropped to 5% with it.73% similarUnverifiedIn manually proofread samples, the character error rate was only 0.68% when the text was clearly legible, but approximately 45% of sampled passages were completely uncorrectable due to source quality issues.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1114267API
curl https://kongchang.com/api/v1/knowledge/claims/1114267MCP
get_claim(id=1114267)