ClaimLens Evidence Map Sample
一个合成声明分析样例，演示 ClaimLens 如何把声明拆成 claim、evidence、source type、uncertainty 和 conclusion boundary。
Static Interactive Sample — Not a live app demo
This is a synthetic claim analysis sample. It is not a live fact-checking system.
Interactive Panels
Click a panel below to explore:
Claim
"Regular exercise reduces the risk of cognitive decline in adults over 60."
A claim is a statement that can be evaluated. It is not a fact until supported.
Evidence
Meta-analysis of 23 longitudinal studies (n = 45,000).
Evidence must be specific: sample size, study design, time horizon, outcome measure.
Source Type
Peer-reviewed journal article, published 2021.
Source type determines trust baseline. Preprint ≠ peer review ≠ systematic review.
Uncertainty
Effect size: moderate. Confidence: medium-high. Confounders: diet, education, baseline health.
Uncertainty is not weakness. It is honest calibration of what we know and what we do not.
Boundary
Applies to adults 60+ in high-income countries. Does not generalize to younger populations or different exercise types.
Every claim has a boundary. ClaimLens makes it explicit.
Next Check
Look for contradictory evidence. Check if the meta-analysis has been replicated or criticized.
A claim is never "done." It is only "checked to this point."
Current confidence: medium — enough evidence to tentatively accept, with noted limitations.
Public Links
How this maps to the project
This sample demonstrates the core structure of claimlens. The real system applies this logic to actual content, with more panels, richer metadata, and export options.
Visit the project page for the full description, and the related Writing / Lab / Case Study for deeper context.