Dataset Viewer
Auto-converted to Parquet Duplicate
adsh
stringlengths
20
20
n_windows
int64
1
8
βŒ€
vec
listlengths
768
768
βŒ€
0000007084-21-000008
4
[ 0.080474295, 0.045906365, -0.18076052, 0.034481615, 0.07563095, -0.0014979035, 0.052296232, -0.0040930575, 0.06099031, 0.011063712, 0.036014643, 0.038397305, 0.057945225, -0.016051663, 0.068156354, 0.03440202, 0.01787695, 0.017926129, -0.004363809, 0.05056025, 0.023561753, ...
0000007084-21-000018
3
[ 0.0851215, 0.032915976, -0.17461768, -0.0103449365, 0.019790089, 0.0011856551, 0.05595603, 0.01195508, 0.044823755, 0.013738379, 0.027234565, 0.033824094, 0.048904702, -0.028461283, 0.058964793, 0.021662517, 0.0133127775, 0.0018216308, -0.009491183, 0.05309929, 0.0088017145, ...
0000007084-21-000031
2
[ 0.06466228, 0.010778816, -0.20364967, 0.014237566, 0.010715823, -0.02669873, 0.040006813, -0.0053326464, 0.020970391, 0.020597657, 0.053009827, 0.012332456, 0.076849714, -0.017272014, 0.08209284, 0.00066393823, 0.009272252, -0.010091915, 0.010747338, 0.08072368, 0.054047853, ...
0000007084-21-000039
3
[ 0.07507593, 0.046715427, -0.2092503, 0.0152449, 0.021645796, 0.029096765, 0.054624524, -0.010937408, 0.041880935, 0.0019602184, 0.04316494, 0.072330765, 0.07216396, -0.016266689, 0.080738835, 0.024298798, 0.029996412, -0.01793083, -0.00019575519, 0.08533228, 0.02842139, 0.0...
0000009346-21-000024
8
[ 0.09981495, 0.0701831, -0.14200984, 0.034197286, 0.07302322, 0.030646909, 0.038940098, 0.04262722, 0.10009312, 0.002183893, 0.042152774, 0.043866333, 0.07727555, 0.022988187, 0.08762392, 0.006679639, 0.023532651, 0.05006148, -0.013673901, 0.041457582, 0.01257282, -0.0233282...
0000009346-21-000037
8
[ 0.1196122, 0.03893598, -0.14882469, 0.029168513, 0.05110863, 0.012651677, 0.05284546, 0.03303818, 0.07013011, 0.019903634, 0.03392679, 0.03806995, 0.090629295, 0.026621249, 0.091694504, 0.020277478, 0.045465793, 0.0032532562, 0.005207654, 0.0847536, 0.026372036, -0.01810307...
0000010254-21-000045
8
[ 0.07940961, 0.08513562, -0.14957476, 0.0232717, 0.057733927, 0.07252265, 0.007296467, 0.08055138, 0.040798087, 0.008301183, 0.08343925, 0.017653849, 0.07281779, 0.010215496, 0.09870956, 0.015812479, 0.010092886, 0.006417714, 0.010826824, 0.0639802, -0.005395977, -0.00241832...
0000010254-21-000074
8
[ 0.035978287, 0.056709398, -0.15242034, -0.0074819596, 0.045670446, 0.01665401, 0.029737547, 0.0744895, 0.035348378, 0.04050695, 0.03560546, 0.027705811, 0.106560946, -0.0068041408, 0.05177809, 0.008150426, 0.070729904, -0.0025624342, 0.001120704, 0.052973624, 0.029523736, -...
0000010254-21-000095
8
[ 0.06004364, 0.035278227, -0.15419486, 0.0062895175, 0.06392716, 0.03261243, 0.049261708, 0.04566557, 0.08253984, 0.025037028, 0.052503638, 0.06081424, 0.06410477, 0.0060729296, 0.08233818, 0.03668397, 0.004346619, 0.026276967, 0.018687624, 0.06603257, 0.010849128, -0.018168...
0000010254-21-000131
8
[ 0.058724493, 0.059776153, -0.17247134, 0.0037281883, 0.05706561, 0.029344112, 0.050147176, 0.045475844, 0.07210063, 0.02200024, 0.05290281, 0.036677074, 0.055927455, 0.0023742379, 0.10783013, 0.04773866, 0.02003725, 0.02670747, 0.024091939, 0.043939438, 0.022977853, -0.0217...
0000010254-21-000167
8
[ 0.065879405, 0.07157333, -0.17368825, -0.015527017, 0.070590965, 0.029609164, 0.057886582, 0.06630811, 0.05828222, 0.019505633, 0.07022577, 0.029456176, 0.052141447, 0.005140386, 0.08562184, 0.06014741, 0.024809834, 0.017528364, 0.04106668, 0.051674206, 0.0121317245, -0.015...
0000015615-21-000006
8
[ 0.1063373, 0.10743593, -0.1632353, 0.016000764, 0.051181808, 0.049425043, 0.050420117, 0.05988388, 0.07630863, 0.023421394, 0.054357633, 0.019487763, 0.048544623, 0.015912518, 0.077314794, 0.0060213846, 0.03191378, 0.03494187, -0.005267412, 0.07082859, 0.012060203, -0.00095...
0000015615-21-000017
1
[ 0.041499127, -0.038868833, -0.17991197, 0.01748122, 0.034654304, 0.007438036, 0.06911852, -0.034174334, 0.028510775, -0.04242862, 0.03640225, 0.017345421, 0.019717477, 0.005511203, 0.03044655, -0.017713886, -0.008976174, -0.0050951135, -0.048889585, 0.027507372, 0.0040743975,...
0000015615-21-000027
1
[ 0.039408635, -0.040057093, -0.18130721, 0.016800357, 0.030743206, 0.005070258, 0.06680003, -0.037868198, 0.029824955, -0.04826575, 0.037886485, 0.0142898, 0.019489562, 0.0075834347, 0.035727162, -0.023418335, -0.005052459, -0.004603827, -0.05097163, 0.027958028, -0.0004167317...
0000015615-21-000035
1
[ 0.02547812, -0.006978821, -0.18271536, 0.020204447, 0.037607525, 0.009783035, 0.06432161, -0.03235942, 0.048866794, -0.032126255, 0.032000314, 0.013060427, 0.011634228, -0.007833568, 0.04124889, -0.0118296975, -0.0055941367, -0.0101730665, -0.051954273, 0.03644106, 0.01387839...
0000019617-21-000236
8
[ 0.097936526, 0.03810124, -0.16139518, 0.01849217, 0.026784806, 0.004964191, 0.08595621, 0.024933841, 0.06378091, 0.036940094, 0.025240203, 0.016062992, 0.068256445, 0.042492144, 0.05189617, 0.021300998, 0.028617859, 0.0026505522, 0.010202819, 0.04188751, 0.015241686, -0.007...
0000019617-21-000292
8
[ 0.09546292, 0.06479379, -0.15320252, 0.015896875, 0.03816035, -0.017034365, 0.07952696, 0.038046174, 0.08079452, 0.022236789, -0.000056832498, 0.021785278, 0.05584032, 0.023959987, 0.055952262, 0.03453374, -0.0030217047, 0.010778323, 0.03566059, 0.037869606, 0.008672929, -0...
0000019617-21-000373
8
[ 0.0847395, 0.06314694, -0.1636251, 0.014482971, 0.050908618, -0.0002147136, 0.08017798, 0.028773423, 0.04797992, 0.006710316, -0.006695328, 0.04657775, 0.06880306, 0.027136266, 0.073430896, 0.03227338, 0.020792427, 0.028061453, 0.013501871, 0.06809345, 0.024621684, -0.02197...
0000019617-21-000458
8
[ 0.086661875, 0.07061813, -0.16503078, 0.013854754, 0.030477926, -0.0013767077, 0.08341092, 0.018537622, 0.07209591, 0.0063407007, -0.007672479, 0.020058252, 0.075608686, 0.0027672986, 0.07749773, 0.036243387, -0.0002569772, 0.015745185, 0.013854665, 0.06263867, 0.02056149, ...
0000022444-21-000012
8
[ 0.083050855, 0.054801136, -0.17272282, 0.018714486, 0.057027195, 0.011616132, 0.08021647, -0.0011984024, 0.050420422, 0.03548338, 0.022241931, 0.014549307, 0.10478024, -0.0033462155, 0.11576479, 0.05235483, 0.03287961, 0.036617287, 0.03478887, 0.08207846, 0.03303567, -0.003...
0000022444-21-000058
8
[ 0.101426, 0.07227306, -0.16626205, 0.009025064, 0.057532225, -0.011461449, 0.080387436, 0.006733998, 0.04713363, 0.060230173, 0.021778932, 0.03953374, 0.09890183, 0.022169495, 0.10928489, 0.037892792, 0.009023, 0.04093357, 0.040727425, 0.09468763, 0.040656455, 0.012651059, ...
End of preview. Expand in Data Studio

SEC Restatement Corpus (8-K Item 4.02)

156,732 SEC filings from 5,085 companies, 2004 to 2024, labelled for whether the filing was later repudiated by the company that filed it.

Labels are mechanical, not annotated. When a registrant decides previously issued financials should no longer be relied upon it files an 8-K under Item 4.02, and that 8-K names the filings it retracts. The label is that join.

Files

file rows what it is
filings.parquet 156,732 the labelled index: cik, adsh, form, period, filed, sic, label
views_2021.parquet ~4.3k the 2021 subset that has both a narrative and XBRL facts
narratives.parquet 6,184 extracted narrative sections
xbrl_facts.parquet β€” reported tags and values per filing
embeddings_nomic_embed_text_mean.parquet β€” 768-dim, mean pooled over 400-word windows
embeddings_nomic_embed_text_max.parquet β€” the same, max pooled

Two things to know before using it

The join is not positional. One 8-K routinely repudiates a run of periods. Matching by position measured 91.5% accurate and was 67.1% when verified against the 8-K prose. This corpus is built from the prose-named periods.

One industry code carries the 2021 spike. SIC 6770 blank-check shells are 25% of 2021 rows and 48% of its positives, retracted at 0.868 against 0.328 elsewhere. A model given the industry code will identify which industry had the problem and it will look like skill.

Provenance and limits

Source documents are public EDGAR filings, which are US government works. Coverage by filing year is uneven, so a rate is comparable across years and a raw count is not. Recent filings have had less time to be retracted, so their rate is understated rather than over.

Code, results and the paper: https://github.com/hmmrlabs/cost-functions

Downloads last month
98