Datasets:
id stringlengths 23 23 | text stringlengths 599 3.9k | task stringclasses 2
values | phase stringclasses 1
value | generator stringclasses 1
value | group_id stringlengths 17 21 | prompt_chars int64 344 2.67k | n_tokens int64 218 1.5k | text_sha256 stringlengths 64 64 | meta stringlengths 645 2.87k |
|---|---|---|---|---|---|---|---|---|---|
tables-fe960695ce5c3306 | <|start|><|user|>
Please put the following table back into proper rows and columns as a GitHub-style Markdown table. Put the table title on its own line(s) above the table. Combine the two header levels into one header per column, written as "<group>: <subheading>" (for example "Complaints: Opened"). Footnote markers s... | table_md | stable | deterministic | tables|20261019|6121 | 1,002 | 675 | 653f3a6fbb67bbca8826557d2bf27e7f4657efb8a228ee9dc23a05d44c57345b | {"answer_form": "md", "cells": [["Year", "Complaints: Opened", "Complaints: Closed", "Complaints: Pending"], ["1998", "—", "3,101", "45,032"], ["1999", "21,618", "7,705", "—"], ["2000", "44,065", "43,207", "4,563"], ["2001", "41,627", "15,906", "6,510"], ["2002*", "38,307", "29,101", "5,957"], ["2003", "6,869", "23,187... |
tables-ef1676b5d148654b | Please put the following table back into proper rows and columns as a Markdown pipe table. Put the table title on its own line(s) above the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Footnote markers stay where they are; put the footnote t... | table_md | stable | deterministic | tables|20261019|2904 | 1,033 | 454 | 959aecf5aa19bba7c090fd18978f04542288d2f4e65925b259e979726c729411 | {"answer_form": "md", "cells": [["Service", "Fee", "Effective date"], ["Copy of court record, per page", "$150.00", "8/1/2007"], ["Exemplification of a document", "$0.25 per page*", "12/1/2016"], ["Electronic copy on portable media", "$85.00 plus $2.00 per page", "8/1/2020"], ["Record retrieval from storage", "$150.00"... |
tables-9d3d0e44b9b45646 | <|start|><|user|>
Rebuild the original table from this text extraction and give it as a Markdown table. Put the table title on its own line(s) above the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Section-label rows get empty value cells. C... | table_md | stable | deterministic | tables|20261019|5539 | 1,978 | 962 | 10ee2361d19c48ca111c90321898819f4730e94a06ae19f92f9226e04e255984 | {"answer_form": "md", "cells": [["", "Year Ended December 31 / 2001", "Year Ended December 31 / 2000"], ["Revenues:", "", ""], ["Product revenue", "54,918,902", "1,442,138"], ["Service revenue", "15,788,998", "964,762"], ["Total revenues", "70,707,900", "2,406,900"], ["Cost of revenues", "43,878,275", "1,391,610"], ["G... |
tables-b29b7d8ef4e93bac | Table reconstruction: turn this PDF text-layer table into a Markdown pipe table with the same rows and columns. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234").
Text:
Schedule E—Rates
Usa... | table_md | stable | deterministic | tables|20261019|612 | 688 | 798 | cab98a885f69bc593fb7c1f370fd73cb4eb1f2e0f853e275bded19da3809a090 | {"answer_form": "md", "cells": [["Usage tier (kWh)", "Tier 1", "Tier 2", "Tier 3", "Tier 4"], ["1–500", "$5.00", "$5.75", "$6.50", "$7.26"], ["501–1,000", "$5.50", "$6.33", "$7.16", "$7.98"], ["1,001–1,500", "$6.00", "$6.91", "$7.81", "$8.71"], ["1,501–2,000", "$6.50", "$7.48", "$8.46", "$9.43"], ["2,001–2,500", "$7.01... |
tables-d24238d2890990f3 | <|start|><|user|>
Reconstruct the table below, which was copied from a PDF, as a Markdown table. Put the table title on its own line(s) above the table. Rows that are only a section label (such as "Revenues:") keep empty cells in the other columns. The first column heading is empty.
<text>
Yarrowdale Sy... | table_md | stable | deterministic | tables|20261019|7317 | 1,413 | 888 | e94b293f8c0be22ace233a9f3cb54b54eae28fa857d2511e52bdc9c39c4fcaa4 | {"answer_form": "md", "cells": [["", "2008", "2007"], ["Current assets:", "", ""], ["Cash and cash equivalents", "2,734,360", "2,651,910"], ["Accounts receivable, net", "2,252,740", "2,903,090"], ["Inventories", "1,380,680", "723,580"], ["Total current assets", "6,367,780", "6,278,580"], ["Property and equipment, net",... |
tables-fd6ed6fbda3c21c0 | Please put the following table back into proper rows and columns as a Markdown table. Put the table title on its own line(s) above the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Section-label rows get empty value cells. The table has group... | table_md | stable | deterministic | tables|20261019|5725 | 1,699 | 975 | b30c28598cd4604d8d073512a3edcef00c06de2354f82f6c86794992dc52c558 | {"answer_form": "md", "cells": [["", "Fiscal Year - 2020", "Fiscal Year - 2019"], ["Revenues:", "", ""], ["Product revenue", "5,025,081", "23,296,496"], ["Service revenue", "3,028,019", "14,529,104"], ["Total revenues", "8,053,100", "37,825,600"], ["Cost of revenues", "3,109,627", "19,045,688"], ["Gross profit", "4,943... |
tables-d3a4d71dd10b454a | <|start|><|system|>
You are a careful assistant. Follow the user's instructions exactly.<|eot|>
<|user|>
This table lost its structure when the PDF was converted to text. Rebuild it as a GitHub-style Markdown table, one row per table row and one column per table column. Start with the title lines, then a blank line, th... | table_md | stable | deterministic | tables|20261019|3512 | 1,224 | 828 | 190918ecb52655a3b1ebcefd806c2c99b926fd88cbdfce9536d95855f0090a03 | {"answer_form": "md", "cells": [["", "2023", "December 31, 2022"], ["Current assets:", "", ""], ["Cash and cash equivalents", "$139,375", "$9,651"], ["Accounts receivable, net", "210,988", "219,506"], ["Inventories", "128,685", "164,878"], ["Total current assets", "479,048", "394,035"], ["Property and equipment, net", ... |
tables-7ecf4e80deab1154 | Task: Convert the extracted table to a GitHub-style Markdown table. Keep every cell's text exactly as printed. Start with the title lines, then a blank line, then the table.
Excerpt:
Norwick Authority — appropriations summary
[In thousands of dollars]
FY 2021 ... | table_md | stable | deterministic | tables|20261019|9161 | 776 | 474 | 14d8dc746e92391b43c9ce3911979b4db88175f746a55b421fb1557dfaffcb8e | {"answer_form": "md", "cells": [["Program", "FY 2021 enacted", "FY 2022 request", "+/−"], ["Research and evaluation", "400,433", "528,937", "+128,504"], ["Facilities maintenance", "725,959", "621,965", "−103,994"], ["Program administration", "469,184", "574,775", "+105,591"], ["Equipment", "20,115", "23,802", "+3,687"]... |
tables-5d2b085bf234859c | Turn the text below into a clean Markdown table that matches the printed table cell for cell. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Section-label rows get empty value cells. The ... | table_md | stable | deterministic | tables|20261019|4378 | 1,844 | 948 | 2ab70eb0a8ad997ad8ff9cb8e041af4087ddcb7e215bdd7b71605f1e42f9bb4d | {"answer_form": "md", "cells": [["", "2009", "December 31, 2008"], ["Current assets:", "", ""], ["Cash and cash equivalents", "$1,551,590", "$2,277,170"], ["Accounts receivable, net", "2,699,390", "2,469,260"], ["Inventories", "1,826,670", "317,330"], ["Total current assets", "6,077,650", "5,063,760"], ["Property and e... |
tables-ddb7b93710a4a3ef | Instruction: Reconstruct the table below, which was copied from a PDF, as a Markdown pipe table. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. The table has grouped column headings: join each group... | table_md | stable | deterministic | tables|20261019|8129 | 1,267 | 946 | d8fa8e77c7b5670050b4f910eda3564d1c621b0cef6b57055cfffa631d919560 | {"answer_form": "md", "cells": [["District", "Cases filed / Civil", "Cases filed / Criminal", "Applications / Received", "Applications / Approved", "Applications / Denied"], ["Brindle Point", "8,942", "38,184", "11,658", "40,746", "6,833"], ["Norwick", "40,353", "41,417", "31,685", "12,451", "48,837"], ["Harlan Ridge",... |
tables-ed826dfc02726be2 | <|start|><|user|>
Here is a table pasted from a PDF. Return it as a GitHub-style Markdown table; do not change, round or reformat any value. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234").... | table_md | stable | deterministic | tables|20261019|2959 | 1,779 | 837 | 063d6f110075d4026f94cacdf617c24656d20ac44efdec76e4967f56068aeaab | {"answer_form": "md", "cells": [["", "Fiscal Year - 2022", "Fiscal Year - 2021"], ["Revenues:", "", ""], ["Product revenue", "423,563", "88,541"], ["Service revenue", "161,405", "24,537"], ["Total revenues", "584,968", "113,078"], ["Cost of revenues", "211,730", "74,816"], ["Gross profit", "373,238", "38,262"], ["Opera... |
tables-604be3a59acf691d | <|start|><|user|>
Turn the text below into a clean Markdown table that matches the printed table cell for cell. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Rows that are only a section... | table_md | stable | deterministic | tables|20261019|1104 | 2,053 | 1,082 | e4112b4e0a368871c2c727020105ebea1ffe4c2959d7089733bb8b1d010f3f43 | {"answer_form": "md", "cells": [["", "Year Ended December 31 — 2024", "Year Ended December 31 — 2023"], ["Revenues:", "", ""], ["Product revenue", "$4,434,646", "$4,225,590"], ["Service revenue", "1,538,914", "3,687,720"], ["Total revenues", "$5,973,560", "$7,913,310"], ["Cost of revenues (1)", "3,929,383", "4,473,181"... |
tables-c6c98601189da854 | <|start|><|system|>
You are a helpful assistant.<|eot|>
<|user|>
Table reconstruction: turn this PDF text-layer table into a Markdown table with the same rows and columns. Put the table title on its own line(s) above the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number i... | table_md | stable | deterministic | tables|20261019|12116 | 2,020 | 927 | 7ca507ef673db178fba244177c19bb48c7b5296c2198e906254b6226146d8729 | {"answer_form": "md", "cells": [["", "December 31, 2019", "December 31, 2018"], ["Current assets:", "", ""], ["Cash and cash equivalents", "$744,100", "$1,164,830"], ["Accounts receivable, net", "2,374,380", "902,590"], ["Inventories", "499,290", "1,824,500"], ["Total current assets", "3,617,770", "3,891,920"], ["Prope... |
tables-3da451fd9edc7a08 | ### Task
Reconstruct the table below, which was copied from a PDF, as a GitHub-style Markdown table. Put the table title on its own line(s) above the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. Section-label rows get empty value cells. The first colum... | table_md | stable | deterministic | tables|20261019|8360 | 1,409 | 889 | 1833e633a8fa2f335efdc7f5e50ca34165e9ebc690c16027bbb1265301518df6 | {"answer_form": "md", "cells": [["", "2021", "2020"], ["Revenues:", "", ""], ["Product revenue", "2,906,791", "1,569,594"], ["Service revenue", "2,363,099", "454,756"], ["Total revenues", "5,269,890", "2,024,350"], ["Cost of revenues", "2,904,448", "1,174,770"], ["Gross profit", "2,365,442", "849,580"], ["Operating exp... |
tables-784965d7d8499095 | <|start|><|system|>
You are a careful assistant. Follow the user's instructions exactly.<|eot|>
<|user|>
Convert the extracted table to a GitHub-style Markdown table. Keep every cell's text exactly as printed. Put the table title on its own line(s) above the table. Keep numbers exactly as printed, including parentheses... | table_md | stable | deterministic | tables|20261019|2842 | 689 | 257 | ebd0ca05b9b2d33132f0fbf9e2729ea53a3ce240e048bc3afc8aedf38a03c5e4 | {"answer_form": "md", "cells": [["Service", "Fee"], ["Duplicate license", "$20.50"], ["Appeal to the board", "No charge"], ["Certificate of good standing", "$0.25 per page"], ["Inspection, follow-up", "$200.50"], ["Filing a notice of appeal", "$100.00"]], "footnotes": false, "frame_id": "tab:2", "gold": "Schedule of Fe... |
tables-6e86fe6ce5fe60cd | <|start|><|user|>
Here is a table pasted from a PDF. Return it as a Markdown table; do not change, round or reformat any value. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. Section-label rows get ... | table_md | stable | deterministic | tables|20261019|11950 | 2,118 | 1,238 | b896ae76c964f968d815e903b2e5d6ba7f333d41d087061e2295a1d8ec1d72cb | {"answer_form": "md", "cells": [["", "Fiscal Year — 2022", "Fiscal Year — 2021", "Fiscal Year — 2020"], ["Revenues:", "", "", ""], ["Product revenue", "$2,483,954", "$4,630,911", "$1,175,034"], ["Service revenue", "821,076", "1,500,569", "1,108,036"], ["Total revenues", "$3,305,030", "$6,131,480", "$2,283,070"], ["Cost... |
tables-2505673097a5e87d | [Task]
Reconstruct the table below, which was copied from a PDF, as a CSV table. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Footnote markers stay where they are; put the footnote text... | table_csv | stable | deterministic | tables|20261019|10288 | 1,763 | 1,006 | 5a77e2144e42f4f2c3465edf96371f68919b3a88918533eccbeefe3173533829 | {"answer_form": "csv", "cells": [["Service", "Fee", "Effective date", "Authority"], ["Annual registration", "$30.50", "6/1/2015", "Sec. 19.23"], ["Search of records, per name or item", "$40.50", "7/1/2009", "Sec. 23.3"], ["Permit application review", "$1.00 per page", "8/1/2005", "Sec. 24.22"], ["Filing a notice of app... |
tables-90a41f4020e2fc08 | Turn the text below into a clean CSV table that matches the printed table cell for cell. Put the table title on its own line(s) above the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. Rows that are only a section label (such as "Revenues:") keep empty c... | table_csv | stable | deterministic | tables|20261019|5646 | 1,982 | 1,092 | 282317dc51bfdd2c0d08c2aec97f72d3f2f6dafa81c973e7e59b0bdd8c4c5b81 | {"answer_form": "csv", "cells": [["", "Year Ended December 31 / 1999", "Year Ended December 31 / 1998", "Year Ended December 31 / 1997"], ["Revenues:", "", "", ""], ["Product revenue", "$396,868", "$309,031", "$98,800"], ["Service revenue", "385,558", "135,387", "26,025"], ["Total revenues", "$782,426", "$444,418", "$1... |
tables-d0a9da0c91fc62f6 | Query: Table reconstruction: turn this PDF text-layer table into a GitHub-style Markdown table with the same rows and columns. Start with the title lines, then a blank line, then the table.
<text>
Copper Mesa Authority — appropriations summary
(dollars in thousands)
... | table_md | stable | deterministic | tables|20261019|5679 | 772 | 445 | fa9e3467684748ef7d928e80edbeb18d63a71ad7359ba77fe811a272245496a5 | {"answer_form": "md", "cells": [["Program", "FY 2020 enacted", "FY 2021 request", "+/−"], ["Grants to States", "810,531", "1,078,073", "+267,542"], ["Rural assistance", "158,580", "179,705", "+21,125"], ["Equipment", "138,860", "148,597", "+9,737"], ["Emergency reserve", "546,534", "487,910", "−58,624"], ["Total", "1,6... |
tables-0de22caaad4db0f7 | Query: Convert the extracted table to a Markdown pipe table. Keep every cell's text exactly as printed. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Rows that are only a section label (... | table_md | stable | deterministic | tables|20261019|10211 | 1,908 | 972 | 97cb6b8e470cbd450fa6a6e21f12a12452ec6de58ca85790fd699a8b6e3da82b | {"answer_form": "md", "cells": [["", "2005", "2004"], ["Revenues:", "", ""], ["Product revenue", "2,710,404", "4,351,676"], ["Service revenue", "922,496", "2,092,054"], ["Total revenues", "3,632,900", "6,443,730"], ["Cost of revenues (1)", "1,955,909", "3,498,299"], ["Gross profit", "1,676,991", "2,945,431"], ["Operati... |
tables-53ee36edc4f206ac | Q: Table reconstruction: turn this PDF text-layer table into a GitHub-style Markdown table with the same rows and columns. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to.
"""
Larkspur Auth... | table_md | stable | deterministic | tables|20261019|405 | 1,033 | 641 | 3aea2ef7db1551bcd46c89e0d9840d997cff5db599c7c2d88a0ef18435bf01a9 | {"answer_form": "md", "cells": [["Program", "FY 2004 enacted", "FY 2005 request", "+/−"], ["Construction", "520,982", "413,925", "−107,057"], ["Equipment", "763,633", "514,968", "−248,665"], ["Facilities maintenance", "401,020", "328,880", "−72,140"], ["Grants to States", "699,599", "737,125", "+37,526"], ["Program adm... |
tables-d61f0a2ba9505e19 | <|start|><|system|>
You are a helpful assistant.<|eot|>
<|user|>
Reconstruct the table below, which was copied from a PDF, as a Markdown table. Put the table title on its own line(s) above the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to.
---
Table 11.... | table_md | stable | deterministic | tables|20261019|6021 | 766 | 315 | fa24298e32899a6acfb75a7ac1e65fec5833cf2c2c422ff10011634ab8b12691 | {"answer_form": "md", "cells": [["Service", "Fee"], ["Audio recording of a hearing", "$20.50"], ["Reinstatement after suspension", "$15.00"], ["Record retrieval from storage", "$0.50 per page"], ["Late renewal penalty", "$0.10 per page"], ["Copy of court record, per page", "$70.00 plus $2.00 per page"], ["Search of rec... |
tables-11d30c5086087b36 | Table reconstruction: turn this PDF text-layer table into a CSV table with the same rows and columns. Put the table title on its own line(s) above the table.
Text:
Hazelmere Planning Commission — Members
Name Position District Term expires
---------------------------------------------... | table_csv | stable | deterministic | tables|20261019|7343 | 1,230 | 724 | d093728aaa77577d3eee491a1212d81fc5b2e555a6d93cc5f037dab61f54cf1c | {"answer_form": "csv", "cells": [["Name", "Position", "District", "Term expires"], ["Kendra N. Merriweather", "Chair", "1", "Sept. 30, 2030"], ["Agatha D. Dunmore", "Vice Chair", "12", "June 30, 2028"], ["Quinn L. Abernathy", "Secretary", "7", "June 30, 2027"], ["Rosalind E. Calloway", "Treasurer", "8", "Jan. 1, 2024"]... |
tables-0dd0058fa35cf1fe | Turn the text below into a clean GitHub-style Markdown table that matches the printed table cell for cell. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Rows that are only a section labe... | table_md | stable | deterministic | tables|20261019|2683 | 1,693 | 895 | 4beef04c027f30436cc3118f577c38d3b64a90bc71783908cb661a2c5fce0355 | {"answer_form": "md", "cells": [["", "2004", "2003"], ["Revenues:", "", ""], ["Product revenue", "$539,608", "$612,150"], ["Service revenue", "202,564", "206,755"], ["Total revenues", "$742,172", "$818,905"], ["Cost of revenues (1)", "419,538", "400,982"], ["Gross profit", "322,634", "417,923"], ["Operating expenses:",... |
tables-b847d2300983acfb | <|start|><|user|>
Please put the following table back into proper rows and columns as a GitHub-style Markdown table. Put the table title on its own line(s) above the table. The table has grouped column headings: join each group heading and its column heading as "Applications - Received".
Text:
Tabl... | table_md | stable | deterministic | tables|20261019|9025 | 1,357 | 973 | 7a2a4e1087966d899193be0e8585890b971042316bfe4479234a5893f8f02cba | {"answer_form": "md", "cells": [["District", "Applications - Received", "Applications - Approved", "Applications - Denied", "Inspections - Scheduled", "Inspections - Completed"], ["Foxglove", "44,870", "13,783", "48,590", "39,083", "—"], ["Larkspur", "16,390", "41,054", "36,985", "42,701", "—"], ["Stillwater Bend", "12... |
tables-95cce4b458ab3b2f | This table lost its structure when the PDF was converted to text. Rebuild it as a CSV table, one row per table row and one column per table column. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1... | table_csv | stable | deterministic | tables|20261019|3641 | 1,624 | 634 | 0655488a475f558b938a72aaf2eb1c4d6ab1cd98932ede65b353f1919832797d | {"answer_form": "csv", "cells": [["Service", "Fee", "Authority"], ["Facility use, half day", "$15.00", "Sec. 31.18"], ["Annual registration", "No charge", "Sec. 32.10"], ["Audio recording of a hearing", "$0.25 per page", "Sec. 29.3"], ["Amendment to articles", "$10.00", "Sec. 39.15"], ["Certification of a document", "$... |
tables-4cf7b1e6b64215af | <|start|><|user|>
This table lost its structure when the PDF was converted to text. Rebuild it as a Markdown table, one row per table row and one column per table column. Start with the title lines, then a blank line, then the table. Rows that are only a section label (such as "Revenues:") keep empty cells in the other... | table_md | stable | deterministic | tables|20261019|3414 | 1,121 | 923 | 0aa421ceb91c728eb24865db1aa3d14a910182bb96a3f687c6bbdd4b20faca35 | {"answer_form": "md", "cells": [["", "2003", "2002"], ["Current assets:", "", ""], ["Cash and cash equivalents", "17,603,700", "19,090,800"], ["Accounts receivable, net", "21,720,800", "20,745,700"], ["Inventories", "19,502,900", "11,021,400"], ["Total current assets", "58,827,400", "50,857,900"], ["Property and equipm... |
tables-dbfefeb2f735ab6c | ### Task
Turn the text below into a clean CSV table that matches the printed table cell for cell. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to.
Text:
Norwick Authority — appropriation... | table_csv | stable | deterministic | tables|20261019|99 | 1,254 | 721 | 31675c45a512771b48b4c4b5837247fa53a7a56c9b9bf31a9e0f72a813851865 | {"answer_form": "csv", "cells": [["Program", "FY 2020 enacted", "FY 2021 request", "+/−"], ["Equipment", "28,904", "22,179", "−6,725"], ["Facilities maintenance", "292,000", "366,007", "+74,007"], ["Outreach", "473,923", "442,877", "−31,046"], ["Training and technical assistance", "75,420", "100,049", "+24,629"], ["Res... |
tables-1053ad07fff71ab5 | Q: Rebuild the original table from this text extraction and give it as a CSV table. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Rows that are only a section label (such as "Revenues:")... | table_csv | stable | deterministic | tables|20261019|1002 | 1,738 | 944 | 13f079f5bd3d4c0b0de2ab720214cc0b08866355d0b3cc9ca45c548d33dfedb7 | {"answer_form": "csv", "cells": [["", "2004", "2003"], ["Revenues:", "", ""], ["Product revenue", "$9,598,507", "$39,051,670"], ["Service revenue", "4,147,993", "13,319,130"], ["Total revenues", "$13,746,500", "$52,370,800"], ["Cost of revenues", "5,708,849", "19,116,529"], ["Gross profit", "8,037,651", "33,254,271"], ... |
tables-30fa879ca2aedf0f | Convert the extracted table to a CSV table. Keep every cell's text exactly as printed. Put the table title on its own line(s) above the table. The table has grouped column headings: join each group heading and its column heading as "Permits issued / Residential".
Passage:
Table 20.—Caseload stati... | table_csv | stable | deterministic | tables|20261019|12275 | 1,073 | 685 | ba9dae999c009cca8f64ceae826a80d6cfa448778407e8bd94ef5ed2dfc91501 | {"answer_form": "csv", "cells": [["Region", "Permits issued / Residential", "Permits issued / Commercial", "Permits issued / Industrial", "Inspections / Scheduled", "Inspections / Completed"], ["Kingsfold", "39,712", "45,849", "2,322", "19,264", "—"], ["Copper Mesa", "29,651", "17,721", "9,837", "31,667", "29,411"], ["... |
tables-142763a5b36fbf7e | <|start|><|user|>
Table reconstruction: turn this PDF text-layer table into a CSV table with the same rows and columns. Start with the title lines, then a blank line, then the table. The table has grouped column headings: join each group heading and its column heading as "Cases filed: Civil". Keep footnote markers insi... | table_csv | stable | deterministic | tables|20261019|13463 | 1,309 | 982 | f6133cc4ab3551b9a5d9024d11ad6b3d9aaa523a024042c540b2330d43705834 | {"answer_form": "csv", "cells": [["Fiscal year", "Cases filed: Civil", "Cases filed: Criminal", "Inspections: Scheduled", "Inspections: Completed"], ["2011", "48,219", "45,840", "41,381", "25,931"], ["2012*", "11,670", "—", "39,531", "29,860"], ["2013", "43,931", "30,894", "29,138", "1,268"], ["2014", "21,052", "7,829"... |
tables-4604e2db210b534c | <|start|><|user|>
Reconstruct the table below, which was copied from a PDF, as a Markdown table. Put the table title on its own line(s) above the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. Rows that are only a section label (such as "Revenues:") keep... | table_md | stable | deterministic | tables|20261019|10182 | 1,901 | 1,140 | 4e27e6e48cea1bcfd1d36950f01371f0cfd150911d09e0fc194a9c8c8a23d03a | {"answer_form": "md", "cells": [["", "1998", "1997", "1996"], ["Revenues:", "", "", ""], ["Product revenue", "516,853", "309,700", "465,231"], ["Service revenue", "317,277", "229,436", "394,072"], ["Total revenues", "834,130", "539,136", "859,303"], ["Cost of revenues (1)", "472,673", "330,217", "517,091"], ["Gross pro... |
tables-7da4c6d2d56932d2 | ### Task
Turn the text below into a clean Markdown table that matches the printed table cell for cell. Start with the title lines, then a blank line, then the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234").
"""
... | table_md | stable | deterministic | tables|20261019|10373 | 1,275 | 1,033 | c3cce7ba22cae8d7ffed14880602719e935666bd799f94526eaa7d509ddfed96 | {"answer_form": "md", "cells": [["Weight not over (lbs.)", "Zone 1", "Zone 2", "Zone 3", "Zone 4", "Zone 5", "Zone 6"], ["1", "$7.46", "$8.58", "$9.70", "$10.82", "$11.94", "$13.06"], ["2", "$8.21", "$9.44", "$10.67", "$11.90", "$13.13", "$14.36"], ["3", "$8.95", "$10.30", "$11.64", "$12.98", "$14.32", "$15.67"], ["4",... |
tables-6289ce21f387664e | Question: Turn the text below into a clean Markdown pipe table that matches the printed table cell for cell. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. Combine the two header levels into one hea... | table_md | stable | deterministic | tables|20261019|7918 | 1,290 | 722 | 701c7c605d8eb300005898e0b0da4a9f31cc7312b061a2558e5d359d02ddeec1 | {"answer_form": "md", "cells": [["County", "Applications — Received", "Applications — Approved", "Applications — Denied", "Inspections — Scheduled", "Inspections — Completed"], ["Eagle Pass Run", "38,133", "46,977", "44,605", "35,103", "45,083"], ["Ashby Crossing", "32,806", "47,636", "15,823", "7,401", "20,088"], ["Ce... |
tables-effda91cb1f4410c | <|start|><|user|>
This table lost its structure when the PDF was converted to text. Rebuild it as a GitHub-style Markdown table, one row per table row and one column per table column. Start with the title lines, then a blank line, then the table. Combine the two header levels into one header per column, written as "<gr... | table_md | stable | deterministic | tables|20261019|7264 | 1,636 | 1,191 | 3e1bbd43375cdec32eacfac2da6d79eb873acedeebca85325445cb337a9b0fea | {"answer_form": "md", "cells": [["Calendar year", "Complaints - Opened", "Complaints - Closed", "Complaints - Pending", "Inspections - Scheduled", "Inspections - Completed"], ["1990", "31,556", "21,850", "25,468", "8,686", "31,425"], ["1991", "15,147", "—", "19,458", "3,736", "30,925"], ["1992", "577", "37,882", "48,86... |
tables-1af73328b3d60ee5 | Query: Table reconstruction: turn this PDF text-layer table into a GitHub-style Markdown table with the same rows and columns. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to. Section-label rows get e... | table_md | stable | deterministic | tables|20261019|11344 | 1,724 | 878 | 5ea124b508161f25ff673e6f0713d864ae85a8f2a8a33a02f574c6751889706e | {"answer_form": "md", "cells": [["", "Fiscal Year: 2001", "Fiscal Year: 2000"], ["Revenues:", "", ""], ["Product revenue", "$110,569", "$429,141"], ["Service revenue", "41,114", "268,817"], ["Total revenues", "$151,683", "$697,958"], ["Cost of revenues", "90,904", "312,106"], ["Gross profit", "60,779", "385,852"], ["Op... |
tables-d01e334b26696b3d | <|start|><|user|>
Please put the following table back into proper rows and columns as a GitHub-style Markdown table. Start with the title lines, then a blank line, then the table. Numbers stay as printed (commas, parentheses, signs); write "$" directly before the number it belongs to.
<text>
Schedul... | table_md | stable | deterministic | tables|20261019|2182 | 1,642 | 943 | 1784307068d238801a146fcc6e432ec7b4b262beed796828c645e3fdf7b93ab1 | {"answer_form": "md", "cells": [["Service", "Fee", "Effective date", "Authority"], ["Expedited processing (24 hours)", "No charge", "7/1/2006", "Sec. 39.9"], ["Inspection, follow-up", "$250.00", "10/1/2022", "Sec. 15.17"], ["Plan review, commercial", "$1.00 per page", "6/1/2018", "Sec. 13.28"], ["Late renewal penalty",... |
tables-2582e246cbeee4a1 | <|start|><|user|>
Rebuild the original table from this text extraction and give it as a Markdown table. Put the table title on its own line(s) above the table.
<text>
Juniper Flats Authority — appropriations summary
(dollars in thousands)
Program ... | table_md | stable | deterministic | tables|20261019|11980 | 1,488 | 838 | 4f355fa3ca3f5733c7f4c6b2089293c9c1391017a022f4b35a43b3bac24956b7 | {"answer_form": "md", "cells": [["Program", "FY 1999 enacted", "FY 2000 request", "Change"], ["Grants to States", "628,575", "690,758", "+62,183"], ["Research and evaluation", "25,056", "27,614", "+2,558"], ["Training and technical assistance", "641,323", "491,000", "−150,323"], ["Construction", "431,038", "491,257", "... |
tables-fc7c3c6002419a11 | <|start|><|system|>
You are a careful assistant. Follow the user's instructions exactly.<|eot|>
<|user|>
Reconstruct the table below, which was copied from a PDF, as a CSV table. Put the table title on its own line(s) above the table. The table has grouped column headings: join each group heading and its column heading... | table_csv | stable | deterministic | tables|20261019|2865 | 1,226 | 817 | ec243182174d2107dedea5d951db51a57c612f2b19357f2b6ee328e384ea0fa9 | {"answer_form": "csv", "cells": [["Region", "Applications - Received", "Applications - Approved", "Applications - Denied"], ["Dovecote", "20,492", "6,721", "41,083"], ["Larkspur", "28,228", "41,376", "1,436"], ["Marlow Heights", "2,938", "25,869", "5,386"], ["Juniper Flats", "11,377", "6,968", "17,424"], ["Brindle Poin... |
tables-c269790b935c238a | This table lost its structure when the PDF was converted to text. Rebuild it as a Markdown pipe table, one row per table row and one column per table column. Put the table title on its own line(s) above the table. The table has grouped column headings: join each group heading and its column heading as "Enrollment — Ful... | table_md | stable | deterministic | tables|20261019|10594 | 559 | 295 | f38ba88bd27e3e1c427ff8f2d8167b94f7a5d1d85ee1f2fb4d0b642ccdb8d81f | {"answer_form": "md", "cells": [["Year", "Enrollment — Full-time", "Enrollment — Part-time"], ["2019", "29,026", "30,095"], ["2020", "36,776", "42,007"], ["2021", "21,332", "12,726"], ["2022", "12,935", "46,017"]], "footnotes": false, "frame_id": "tab:1", "gold": "Table 5.—Workload by office, by year\n\n| Year | Enroll... |
tables-ac6cb2eb59dff8b4 | Rebuild the original table from this text extraction and give it as a Markdown pipe table. Put the table title on its own line(s) above the table. Combine the two header levels into one header per column, written as "<group> - <subheading>" (for example "Complaints - Opened").
"""
Table 31.—... | table_md | stable | deterministic | tables|20261019|13167 | 1,007 | 616 | 4af36f060049f31fe7e6a1bf73d7ea843057f4111897938093002e59e502413d | {"answer_form": "md", "cells": [["Region", "Complaints - Opened", "Complaints - Closed", "Complaints - Pending", "Permits issued - Residential", "Permits issued - Commercial", "Permits issued - Industrial"], ["Stillwater Bend", "16,592", "13,379", "21,187", "6,494", "43,151", "1,904"], ["Wexmoor", "25,468", "15,960", "... |
tables-5058e572db55fb70 | <|start|><|user|>
This table lost its structure when the PDF was converted to text. Rebuild it as a Markdown table, one row per table row and one column per table column. Put the table title on its own line(s) above the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to... | table_md | stable | deterministic | tables|20261019|8851 | 1,769 | 899 | 50d28875e53ed40542ba1ba608b14ad2104f21538b0e76ee1771fefe1d636c47 | {"answer_form": "md", "cells": [["", "2018", "2017"], ["Revenues:", "", ""], ["Product revenue", "$431,051", "$646,723"], ["Service revenue", "278,586", "188,943"], ["Total revenues", "$709,637", "$835,666"], ["Cost of revenues", "413,881", "367,118"], ["Gross profit", "295,756", "468,548"], ["Operating expenses:", "",... |
tables-9b9e5140a9cdf43c | Reconstruct the table below, which was copied from a PDF, as a CSV file (comma-separated, header row first). Put the table title on its own line(s) above the table. Keep numbers exactly as printed, including parentheses for negatives; attach a dollar sign to its number ("$1,234"). Section-label rows get empty value cel... | table_csv | stable | deterministic | tables|20261019|704 | 1,172 | 1,011 | e3e2acb23b44b6ba934ee38549cf6318d80e37dc4ea273c8882e306566668cf6 | {"answer_form": "csv", "cells": [["", "1998", "1997", "1996"], ["Revenues:", "", "", ""], ["Product revenue", "636,679", "499,111", "357,104"], ["Service revenue", "258,501", "374,770", "144,299"], ["Total revenues", "895,180", "873,881", "501,403"], ["Cost of revenues", "520,879", "548,354", "308,981"], ["Gross profit... |
End of preview. Expand in Data Studio
ALEA table layout recovery
Plain-text tables (space-aligned with wrapped cells, merged group headers, dot leaders and detached dollar signs; whitespace-separated rows; or one cell per line) are converted to a Markdown pipe table or CSV, with the title and footnotes kept. Table types: income statements, statistical tables with two-level headers, balance sheets, fee schedules, budgets, rosters, rate tables and meeting schedules.
10,404 documents, 7,517,569 tokens (alea-v2b-sw32k, the ALEA R4 tokenizer, counting each document's start and end tokens).
Part of the ALEA R4 training mix
This dataset contains exactly the documents of this family that ALEA's R4 model trained on in its stable pretraining mixture (second data stage, one pass each). Every row is byte-identical to the trained text; text_sha256 is the SHA-256 of text.
Held-out evaluation items generated by the same code are not part of this release.
Composition
| task | documents | tokens |
|---|---|---|
table_md |
8,081 | 5,861,090 |
table_csv |
2,323 | 1,656,479 |
3,653 documents use ALEA's chat template.
How it was made
Written by deterministic code from typed tables; no language model was used. All entities are fictional and no figure is presented as a real statistic. Each instruction states every convention the answer relies on (numbers as printed, how merged headers are joined, where footnote markers go).
Sources
- Invented typed tables (names and values from fixed word lists and random draws)
Verification
- Each answer is parsed back and equals the typed cells, header included.
- The word multiset of the noisy text (leaders and rules dropped) equals that of the title, headers, cells and footnotes.
- Decontamination: training rows that shared a word 13-gram with a held-out evaluation item of this or any sibling family were removed, and rows overlapping ALEA's evaluation sets were removed or logged according to ALEA's decontamination policy.
- Sampled records were reviewed blind by a separate reviewer who recomputed or re-derived the answers; defect patterns found in review were fixed in the generator, or the affected rows dropped, before the final build.
Columns
| column | content |
|---|---|
id |
row id from the generator (not always unique: a few families reuse an id across variants, and a handful of exact duplicate documents were trained twice and appear twice) |
text |
the document exactly as trained; chat documents use ALEA's chat template, whose control tokens appear as text (`< |
task |
task kind |
phase |
stable (trained in R4's stable pretraining mixture) |
generator |
what wrote the row (deterministic = code only) |
group_id |
source passage / record group; all variants of one group share it |
prompt_chars |
length in characters of the request part of text (0 for plain documents; null where not recorded) |
n_tokens |
tokens under alea-v2b-sw32k, including the document start and end tokens |
text_sha256 |
SHA-256 of text |
meta |
JSON: the generator's structured fields (source, presentation, answer form, gold values, checks) |
Known limitations
- Real-world tables (GPO, EDGAR) are not included in this release.
- The documents are generated and more repetitive than natural text.
License
CC-BY-4.0 for this compilation and the generated text. Public-domain and public-record source text keeps its status; Wikidata is CC0; OpenGloss and SHELF are CC-BY-4.0.
Published by the ALEA Institute.
- Downloads last month
- 18