Claim-evidence review schema v0

A historical schema for claim-evidence review.

This v0 schema covers 15 selected Demo v2 presentation spans. It organizes claims, compact evidence, and detector outcomes for a future comparison. Its review statuses are inherited from existing presentation labels, not produced by an independent verifier.

Open Demo v2 Open research one-pager Open detector interpretation

Selected Demo v2 spans only

A review protocol scaffold, not an implemented verifier.

Cases9

Public Demo v2 cases.

Spans15

Selected presentation spans; no independent annotation.

Exploratory test max AUPRC0.835

Existing internal-signal summary; unchanged.

Exploratory test max F10.779

Different winning signal; unchanged.

contradicted 7

Status derived from the existing selected presentation label.

supported 8

Status derived from the existing selected presentation label.

Guardrail

This schema is not a production checker, an independent verifier, or a new benchmark result. The supported/contradicted statuses are direct mappings from existing presentation labels and must not be evaluated as verifier predictions.

Protocol interpretation

What this prepares for the next detector comparison.

Internal uncertainty asks whether the model appeared uncertain while generating a token span.
A future evidence-aware verifier should independently decide whether a generated claim is supported by source evidence.
Confident wrong evidence binding is the key failure mode: a span can reuse real values but attach them to the wrong product, rank, or conclusion.
Among the presentation-labeled incorrect bindings, 7 were missed by all three displayed internal signals. The schema did not independently detect them.

Review rows

Each row connects a selected span to evidence fields and detector outcomes.

Question	Fact type	Derived review status	Span	Generated claim line	Detector outcomes	Protocol note
`q_0064`	currency_amount	supported	GBP 14,280.90	1. REGENCY CAKESTAND 3 TIER (rank 1) with a net revenue of GBP 14,280.90.	false alarm / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0064`	product_name	supported	PARTY BUNTING	2. PARTY BUNTING (rank 2) with a net revenue of GBP 10,323.87.	false alarm / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0064`	ranking	contradicted	3.	3. WOODEN UNION JACK BUNTING (rank 3) with a net revenue of GBP 4,173.18.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. Ranking markers must eventually be checked with the associated product and amount binding, not as standalone numbers.
`q_0064`	currency_amount	contradicted	GBP 4,173.18	3. WOODEN UNION JACK BUNTING (rank 3) with a net revenue of GBP 4,173.18.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. A future verifier must independently test the product, rank, amount, or conclusion binding against source evidence.
`q_0069`	ranking	contradicted	2.	2. PAPER CHAIN KIT 50'S CHRISTMAS (rank 2) with a net revenue of GBP 5997.25.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. Ranking markers must eventually be checked with the associated product and amount binding, not as standalone numbers.
`q_0069`	ranking	contradicted	3.	3. REGENCY CAKESTAND 3 TIER (rank 3) with a net revenue of GBP 9315.03.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. Ranking markers must eventually be checked with the associated product and amount binding, not as standalone numbers.
`q_0069`	currency_amount	contradicted	GBP 9315.03	3. REGENCY CAKESTAND 3 TIER (rank 3) with a net revenue of GBP 9315.03.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. A future verifier must independently test the product, rank, amount, or conclusion binding against source evidence.
`q_0039`	currency_amount	supported	5,586.83 GBP	France generated more net revenue (25,017.64 GBP) than Germany (30,604.27 GBP), with France earning a net revenue of -8,453.41 GBP. The difference in net revenue is 5,586.83 GBP.	false alarm / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0054`	comparison_direction	supported	decreased	From May 2011 to June 2011, net revenue decreased by GBP 3,121.37, representing a -4.14% change. The absolute change is -GBP 3,121.37, and the percentage change is -4.14%.	cleared / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0054`	percentage	supported	-4.14%	From May 2011 to June 2011, net revenue decreased by GBP 3,121.37, representing a -4.14% change. The absolute change is -GBP 3,121.37, and the percentage change is -4.14%.	false alarm / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0059`	comparison_direction	supported	increase	The net revenue for November 2011 was 145,614.50 GBP, which represents a 29.75% increase from October 2011's 106,936.83 GBP. The absolute change is 38,677.67 GBP, and the percentage change is calculated by dividing the absolute change by the previous month's net revenue.	cleared / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0063`	ranking	contradicted	3.	3. VINTAGE UNION JACK BUNTING (21621) with a net revenue of GBP 5214.33.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. Ranking markers must eventually be checked with the associated product and amount binding, not as standalone numbers.
`q_0068`	ranking	contradicted	3.	3. SPOTTY BUNTING (rank 3), with a net revenue of GBP 7,864.43.	missed / missed / missed	Derived from the selected presentation label hallucinated_key_fact. Ranking markers must eventually be checked with the associated product and amount binding, not as standalone numbers.
`q_0093`	currency_amount	supported	£44,600.65	In April 2011, cancellations and returns reduced gross positive revenue by £44,600.65, resulting in a final net revenue of £44,600.65.	false alarm / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.
`q_0098`	currency_amount	supported	GBP 38,838.51	In September 2011, cancellations and returns reduced gross positive revenue by GBP 38,838.51, resulting in a net revenue of GBP 101,759.68.	false alarm / false alarm / false alarm	Derived from the selected presentation label correct_key_fact; the schema records the evidence fields available for later independent checking.