Unverified50% confidenceFactExact time
The study drew 2,048 test patient-condition pairs from TrialGPT and from question-answering datasets built on MIMIC-III and MIMIC-IV.
1
Sources
50%
Confidence
Long-term
Relevance
10/27/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe MIMIC dataset primarily originates from intensive care settings, and the model's generalizability in broader clinical environments remains to be examined.72% similarUnverifiedTEFM's experiments were validated on datasets in critical domains such as clinical care and safety, with competitive classification accuracy.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1103053API
curl https://kongchang.com/api/v1/knowledge/claims/1103053MCP
get_claim(id=1103053)