Open-source · MIT-licensed · Research prototype

Measure sapience.
Publish the number.

The KST Index is the open leaderboard for the Kari-Sheldon Test: a benchmark for the functional behavioral signatures of personhood in AI systems. Seven cognitive-science-grounded sub-tests. One 0–100 composite, gated by an integrity multiplier. Every score replayable and auditable, though not yet validated. It is not a consciousness test, and it never claims to be.

See the leaderboard → How KST works
Status and corrections (October 2026). This site describes the Kari-Sheldon Test as designed, and much of that design is not implemented: there are no trained or certified raters, no reliability statistic, no DIF or fairness analysis, no IRT scaling or calibration sample, and no expert review (the “expert panel” was not a panel of human experts); every item is sent as a single prompt. Scores come from automated keyword, regular-expression and word-overlap rules. They are placeholders, not validated measurements of sapience or of any other construct. See the correction notice in the KST repository.

The KST Index

KST is designed to probe sapience markers, functional behavioral signatures, not consciousness, sentience, or subjective experience. What a score does and does not mean →

Loading the board…

Markers, not vibes

Seven sub-tests measure metacognition under pressure, recursive theory of mind, practical wisdom, affective reasoning, honest refusal, self-revision, and integration, each with a falsifiability criterion stated in advance.

Integrity is a hard cap

A model that confabulates or deceives cannot ride a high reasoning score to a misleading headline. The HRO integrity multiplier caps the composite; a catastrophic-deception flag hard-caps it at 25.

Every number is replayable

A strict JSON audit envelope per item lets anyone replay a score. Bootstrap confidence intervals, a reproducibility statistic and per-population fairness checks are planned, not implemented; the intervals shown today are placeholders.