Company intelligence
BeliefLens
Measure belief. Govern uncertainty.
1 employees
4 public signals
About BeliefLens
BeliefLens turns unpredictable AI outputs into calibrated, auditable decision states—enabling organizations to measure uncertainty, detect instability and govern AI workflows quantitatively at runtime.
Verified activity
Signals from BeliefLens
4 published signals
Research & Knowledge
BeliefLens conducted a controlled 30-case benchmark demonstrating that the Semantic Evidence Graph (SSEG) detected and localized every injected fault without falsely rejecting healthy runs.
Reported by Matthew Dixon, Ph.D., FRM
Products & Services
BeliefLens developed Verity, an algorithm that constructs semantic evidence graphs to connect claims to supporting sources, measure prompt and output uncertainty, test stability, and show where evidence stops explaining answers.
Reported by Matthew Dixon, Ph.D., FRM
Research & Knowledge
BeliefLens is working on constructing a semantic evidence graph from token probabilities and controlled prompt variations to examine model interpretation, uncertainty, and response stability.
Reported by Matthew Dixon, Ph.D., FRM
Presence & Recognition
BeliefLens is hosting a practical AI Quant Risk Bootcamp on 21 October at the NYC STAC Summit covering prompt stability, calibration, stochastic evidence graphs, and open-weight models.
Reported by Matthew Dixon, Ph.D., FRM