Row 1845: page

Source · scan and Parse blocks

Parse text by block 1 blocks reading order

0
PictureXVII.

Provenance page_id, checksum, Parse model

page_id: britannica:309f765b1ecc83e0a7f7762e44b4a927a7555443:train:1845
sha256: a758ad9d45a7173d220cf1ae4e2c98e462082a9ad3b0185ff4d8121b5631b0fa
Parse: nvidia/nemotron-3.5-parse · finish stop

LaTeX from Parse blocks ok Open PDF LaTeX source

Inspector · generated

Page match

plate page

uncertainno usable text Unguía Fault

caption "XVII." only

Scores and reasoning

[W:57161345:1294556424]

Running head topic XVII.

Rerank: -2.32
Margin: 0.25
BM25 rank: 1

supervl+bm25+embedvl+rerankvl

Plate result from artifacts/plates/1845.json: Rerank VL and Embed VL disagree and the logit is below tau

Plate · visual match

uncertain Unguía Fault rev 1294556424

Rerank VL and Embed VL disagree and the logit is below tau

Super VL subject Colombia map

Neighbor scans 2

Source · neighbor scans in the dataset (same volume id). They are not always the next printed pages.

-2 · 1843 CONIC SECTIONS.

-1 · 1844 100

Super VL description

nvidia/Nemotron-3.5-Super-VL description from the plate image and neighbor text

The image displays a very faint, aged engraved plate, likely a map, with minimal visible detail due to fading. A small rectangular label in the lower right corner contains the text 'COLOMBIA MAP'.

Subject: Colombia map
Search: Colombia map; Colombia map 19th century; Colombia geography map; Encyclopaedia Britannica Colombia plate; Colombia South America map

row 1844 COLOMBIA found in source

Ranking per stage 10 candidates

BM25: Unguía Fault
Embed VL: Joaquín Acosta
Rerank VL: Unguía Fault image + caption
Rerank VL: Unguía Fault image only

Candidate	BM25	cos	logit	img
Unguía Fault	1	0.272	-2.32	-2.46
Geographic Institute Agustín Codazzi	2	0.292	-2.58	-2.48
Geography of Cesar Department	5	0.263	-2.91	-2.87
General Archive of the Nation (Colombia)	2	0.274	-3.36	-3.21
Colombian Geological Survey	5	0.256	-3.68	-3.51
Typography (cartography)	4	0.270	-3.75	-3.61
Map tree frog	3	0.296	-3.97	-3.81
Joaquín Acosta	4	0.326	-4.20	-4.17
Outline of South America	4	0.270	-4.34	-4.32
Anarchism in Colombia	3	0.265	-4.91	-4.69

30 BM25 candidates, top 10 by Embed VL cosine sent to Rerank VL. Cosine is document-to-document (plate image vs article text).

Entries · Wikipedia matches

0 matched · 1 uncertain

(continuation)B:1845:0 uncertainno usable text

XVII.

caption "XVII." only

Best candidate: XVII Reserve Corps (German Empire)

Rerank
-1.47

Margin
0.36

Cosine
0.318

BM25 rank
5

bm25+embed+rerank

entry has 5 chars (< 40); too short to judge

Built 2026-10-18 17:02. Britannica scans: public domain (CC0). Wikipedia text: CC BY-SA 4.0, linked by page and revision. OCR text is Parse 3.5 output without correction.