The VERITY encoder measures meaning. Not a similarity engine. A measurement engine. The difference is architectural, and it cannot be closed by scale, data, or fine-tuning.
"Every other system encodes text into a position. VERITY encodes text into a measurement. Position is the output. Fidelity is the purpose."
Traditional embedding systems compress meaning into a single similarity score. That compression is a design choice, and it has consequences. Contrasts that change the claim (a denial, a swapped actor, a weaker commitment, a broader or narrower statement) get blurred together. The result is a system that cannot distinguish "the drug is effective" from "the drug is not effective" because both sentences contain the same words, and words are what similarity systems measure.
VERITY was built around a different premise: meaning is not one thing. It is the agreement and disagreement of multiple independent measurement perspectives, and that disagreement is not noise to hide. It is information. When perspectives conflict, that conflict is more informative than any single confidence score. The encoder was built to report that conflict, not collapse it.
The encoder measures text across independent perspectives simultaneously. When they agree, confidence is high. When they disagree, the system flags the conflict and says so.
The Semantic Fidelity Benchmark tests whether an embedding system can tell when a claim actually changes: opposites, swapped actors, and broader or narrower statements. Those are the discriminations similarity-based systems structurally struggle to make. An expanded 500-pair benchmark with full competitive results is in progress.
On the standard human-aligned similarity benchmark, which neither system was built to optimize for.
An interactive tool for measuring fidelity between any two sentences is coming soon. Compare a claim with its opposite, swap who acted, change how strong or how broad the claim is, and see the discrimination live.