Dataset Viewer
Auto-converted to Parquet Duplicate
id
string
task
string
source
dict
table
dict
segment_facts
dict
segments
dict
question
string
question_language
string
expected_answers
list
answer_comparison
string
inputs
dict
lang_table_01
filtered_maximum
{ "question_id": "nt-0", "question_file": "data/training.tsv", "original_question": "what was the last year where this team was a part of the usl a-league?", "original_answers": [ "2004" ], "table_name": "csv/204-csv/590.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiTab...
{ "name": "csv/204-csv/590.tsv", "header": [ "Year", "League" ], "rows": [ [ "2001", "USL A-League" ], [ "2002", "USL A-League" ], [ "2003", "USL A-League" ], [ "2004", "USL A-League" ], [ "2005", "USL First ...
{ "S1": { "header": [ "Record ID", "Year" ], "rows": [ [ "R401", "2001" ], [ "R123", "2002" ], [ "R676", "2003" ], [ "R162", "2009" ], [ "R937", "2004" ],...
{ "S1": { "A": "For record R401, the year is 2001. For record R123, the year is 2002. For record R676, the year is 2003. For record R162, the year is 2009. For record R937, the year is 2004. For record R604, the year is 2005. For record R759, the year is 2006. For record R424, the year is 2007. For record R682, t...
Among the supplied records, what was the last year this team played in the USL A-League?
en
[ "2004" ]
single_string
{ "AA": "For record R401, the year is 2001. For record R123, the year is 2002. For record R676, the year is 2003. For record R162, the year is 2009. For record R937, the year is 2004. For record R604, the year is 2005. For record R759, the year is 2006. For record R424, the year is 2007. For record R682, the year is ...
lang_table_02
difference
{ "question_id": "nt-3", "question_file": "data/training.tsv", "original_question": "how many more passengers flew to los angeles than to saskatoon from manzanillo airport in 2013?", "original_answers": [ "12,467" ], "table_name": "csv/203-csv/515.csv", "normalized_table_url": "https://raw.githubuserc...
{ "name": "csv/203-csv/515.tsv", "header": [ "City", "Passengers" ], "rows": [ [ "United States, Los Angeles", "14,749" ], [ "United States, Houston", "5,465" ], [ "Canada, Calgary", "3,761" ], [ "Canada, Saskatoon", "2,282" ...
{ "S1": { "header": [ "Record ID", "City" ], "rows": [ [ "R951", "Canada, Calgary" ], [ "R316", "United States, Los Angeles" ], [ "R423", "Canada, Saskatoon" ], [ "R590", "United States, H...
{ "S1": { "A": "For record R951, the destination is Canada, Calgary. For record R316, the destination is United States, Los Angeles. For record R423, the destination is Canada, Saskatoon. For record R590, the destination is United States, Houston.", "B": "| Record ID | City |\n| --- | --- |\n| R951 | Canada, ...
According to the supplied records, how many more passengers flew to Los Angeles than to Saskatoon?
en
[ "12,467" ]
single_string
{ "AA": "For record R951, the destination is Canada, Calgary. For record R316, the destination is United States, Los Angeles. For record R423, the destination is Canada, Saskatoon. For record R590, the destination is United States, Houston.\n\nFor record R316, the passenger count is 14,749. For record R423, the passe...
lang_table_03
pairwise_comparison
{ "question_id": "nt-7", "question_file": "data/training.tsv", "original_question": "which is deeper, lake tuz or lake palas tuzla?", "original_answers": [ "Lake Palas Tuzla" ], "table_name": "csv/204-csv/341.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiTableQuestions/...
{ "name": "csv/204-csv/341.tsv", "header": [ "Name in English", "Depth" ], "rows": [ [ "Lake Van", "451 m" ], [ "Lake Tuz", "2 m" ], [ "Lake Beyşehir", "10 m" ], [ "Lake Palas Tuzla", "15 m" ] ] }
{ "S1": { "header": [ "Record ID", "Name in English" ], "rows": [ [ "R872", "Lake Van" ], [ "R700", "Lake Beyşehir" ], [ "R392", "Lake Tuz" ], [ "R476", "Lake Palas Tuzla" ] ] ...
{ "S1": { "A": "Record R872 describes Lake Van. Record R700 describes Lake Beyşehir. Record R392 describes Lake Tuz. Record R476 describes Lake Palas Tuzla.", "B": "| Record ID | Name in English |\n| --- | --- |\n| R872 | Lake Van |\n| R700 | Lake Beyşehir |\n| R392 | Lake Tuz |\n| R476 | Lake Palas Tuzla |" ...
According to the supplied records, which is deeper, Lake Tuz or Lake Palas Tuzla?
en
[ "Lake Palas Tuzla" ]
single_string
{ "AA": "Record R872 describes Lake Van. Record R700 describes Lake Beyşehir. Record R392 describes Lake Tuz. Record R476 describes Lake Palas Tuzla.\n\nFor record R872, the depth is 451 m. For record R476, the depth is 15 m. For record R392, the depth is 2 m. For record R700, the depth is 10 m.\n\nQuestion: Accordin...
lang_table_04
matching_attribute
{ "question_id": "nt-9", "question_file": "data/training.tsv", "original_question": "which players played the same position as ardo kreek?", "original_answers": [ "Siim Ennemuist", "Andri Aganits" ], "table_name": "csv/203-csv/116.csv", "normalized_table_url": "https://raw.githubusercontent.com/pp...
{ "name": "csv/203-csv/116.tsv", "header": [ "Player", "Position" ], "rows": [ [ "Ardo Kreek", "Middle blocker" ], [ "Kert Toobal", "Setter" ], [ "Martti Juhkami", "Spiker" ], [ "Siim Ennemuist", "Middle blocker" ], [ ...
{ "S1": { "header": [ "Record ID", "Player" ], "rows": [ [ "R841", "Kert Toobal" ], [ "R516", "Ardo Kreek" ], [ "R246", "Martti Juhkami" ], [ "R283", "Andri Aganits" ], [ ...
{ "S1": { "A": "For record R841, the player is Kert Toobal. For record R516, the player is Ardo Kreek. For record R246, the player is Martti Juhkami. For record R283, the player is Andri Aganits. For record R450, the player is Siim Ennemuist.", "B": "| Record ID | Player |\n| --- | --- |\n| R841 | Kert Toobal...
Among the supplied records, which other players have the same position as Ardo Kreek?
en
[ "Siim Ennemuist", "Andri Aganits" ]
unordered_strings
{ "AA": "For record R841, the player is Kert Toobal. For record R516, the player is Ardo Kreek. For record R246, the player is Martti Juhkami. For record R283, the player is Andri Aganits. For record R450, the player is Siim Ennemuist.\n\nFor record R283, the playing position is Middle Blocker. For record R246, the p...
lang_table_05
reverse_lookup
{ "question_id": "nt-26", "question_file": "data/training.tsv", "original_question": "what club scored the only a total of 79 points?", "original_answers": [ "Málaga CF" ], "table_name": "csv/204-csv/256.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiTableQuestions/maste...
{ "name": "csv/204-csv/256.tsv", "header": [ "Club", "Points" ], "rows": [ [ "Málaga CF", "79" ], [ "CD Numancia", "73" ], [ "Sevilla FC", "71" ], [ "UD Las Palmas", "68" ] ] }
{ "S1": { "header": [ "Record ID", "Club" ], "rows": [ [ "R631", "CD Numancia" ], [ "R625", "Málaga CF" ], [ "R738", "Sevilla FC" ], [ "R832", "UD Las Palmas" ] ] }, "S2": { ...
{ "S1": { "A": "For record R631, the club is CD Numancia. For record R625, the club is Málaga CF. For record R738, the club is Sevilla FC. For record R832, the club is UD Las Palmas.", "B": "| Record ID | Club |\n| --- | --- |\n| R631 | CD Numancia |\n| R625 | Málaga CF |\n| R738 | Sevilla FC |\n| R832 | UD L...
Among the supplied records, which club has 79 points?
en
[ "Málaga CF" ]
single_string
{ "AA": "For record R631, the club is CD Numancia. For record R625, the club is Málaga CF. For record R738, the club is Sevilla FC. For record R832, the club is UD Las Palmas.\n\nFor record R625, the points total is 79. For record R631, the points total is 73. For record R832, the points total is 68. For record R738,...
lang_table_06
filter_with_exclusion
{ "question_id": "nt-35", "question_file": "data/training.tsv", "original_question": "who is the other person who is 24 years old besides reyna royo?", "original_answers": [ "Marisela Moreno Montero" ], "table_name": "csv/204-csv/144.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppa...
{ "name": "csv/204-csv/144.tsv", "header": [ "Contestant", "Age" ], "rows": [ [ "Mildred Kincaid Adames", "20" ], [ "Reyna Royo", "24" ], [ "Karol Guevara", "22" ], [ "Marisela Moreno Montero", "24" ] ] }
{ "S1": { "header": [ "Record ID", "Contestant" ], "rows": [ [ "R438", "Mildred Kincaid Adames" ], [ "R610", "Karol Guevara" ], [ "R674", "Reyna Royo" ], [ "R509", "Marisela Moreno Montero...
{ "S1": { "A": "For record R438, the contestant is Mildred Kincaid Adames. For record R610, the contestant is Karol Guevara. For record R674, the contestant is Reyna Royo. For record R509, the contestant is Marisela Moreno Montero.", "B": "| Record ID | Contestant |\n| --- | --- |\n| R438 | Mildred Kincaid Ad...
Among the supplied records, who besides Reyna Royo is 24 years old?
en
[ "Marisela Moreno Montero" ]
single_string
{ "AA": "For record R438, the contestant is Mildred Kincaid Adames. For record R610, the contestant is Karol Guevara. For record R674, the contestant is Reyna Royo. For record R509, the contestant is Marisela Moreno Montero.\n\nFor record R438, the age is 20 years. For record R674, the age is 24 years. For record R50...
lang_table_07
lookup
{ "question_id": "nt-40", "question_file": "data/training.tsv", "original_question": "what was the number of silver medals won by ukraine?", "original_answers": [ "2" ], "table_name": "csv/203-csv/175.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiTableQuestions/master/c...
{ "name": "csv/203-csv/175.tsv", "header": [ "Nation", "Silver" ], "rows": [ [ "Great Britain (GBR)", "1" ], [ "Brazil (BRA)", "0" ], [ "Ukraine (UKR)", "2" ], [ "Denmark (DEN)", "0" ] ] }
{ "S1": { "header": [ "Record ID", "Nation" ], "rows": [ [ "R128", "Great Britain (GBR)" ], [ "R970", "Denmark (DEN)" ], [ "R479", "Ukraine (UKR)" ], [ "R899", "Brazil (BRA)" ] ]...
{ "S1": { "A": "For record R128, the nation is Great Britain (GBR). For record R970, the nation is Denmark (DEN). For record R479, the nation is Ukraine (UKR). For record R899, the nation is Brazil (BRA).", "B": "| Record ID | Nation |\n| --- | --- |\n| R128 | Great Britain (GBR) |\n| R970 | Denmark (DEN) |\n...
According to the supplied records, how many silver medals did Ukraine win?
en
[ "2" ]
single_string
{ "AA": "For record R128, the nation is Great Britain (GBR). For record R970, the nation is Denmark (DEN). For record R479, the nation is Ukraine (UKR). For record R899, the nation is Brazil (BRA).\n\nFor record R899, the silver medal count is 0. For record R128, the silver medal count is 1. For record R970, the silv...
lang_table_08
pairwise_comparison
{ "question_id": "nt-50", "question_file": "data/training.tsv", "original_question": "who earned more medals--vietnam or indonesia?", "original_answers": [ "Indonesia (INA)" ], "table_name": "csv/204-csv/725.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiTableQuestions/m...
{ "name": "csv/204-csv/725.tsv", "header": [ "Nation", "Total" ], "rows": [ [ "Japan (JPN)", "10" ], [ "Indonesia (INA)", "5" ], [ "Vietnam (VIE)", "2" ], [ "Chinese Taipei (TPE)", "4" ] ] }
{ "S1": { "header": [ "Record ID", "Nation" ], "rows": [ [ "R499", "Vietnam (VIE)" ], [ "R818", "Japan (JPN)" ], [ "R440", "Indonesia (INA)" ], [ "R796", "Chinese Taipei (TPE)" ] ...
{ "S1": { "A": "For record R499, the nation is Vietnam (VIE). For record R818, the nation is Japan (JPN). For record R440, the nation is Indonesia (INA). For record R796, the nation is Chinese Taipei (TPE).", "B": "| Record ID | Nation |\n| --- | --- |\n| R499 | Vietnam (VIE) |\n| R818 | Japan (JPN) |\n| R440...
According to the supplied records, which nation won more total medals, Vietnam or Indonesia?
en
[ "Indonesia (INA)" ]
single_string
{ "AA": "For record R499, the nation is Vietnam (VIE). For record R818, the nation is Japan (JPN). For record R440, the nation is Indonesia (INA). For record R796, the nation is Chinese Taipei (TPE).\n\nFor record R499, the total medal count is 2. For record R796, the total medal count is 4. For record R818, the tota...
lang_table_09
argmax
{ "question_id": "nt-52", "question_file": "data/training.tsv", "original_question": "which club had the most losses?", "original_answers": [ "RC Toulonnais" ], "table_name": "csv/203-csv/322.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiTableQuestions/master/csv/203-cs...
{ "name": "csv/203-csv/322.tsv", "header": [ "Club", "Lost" ], "rows": [ [ "Biarritz Olympique", "7" ], [ "SU Agen", "11" ], [ "Montpellier RC", "17" ], [ "RC Toulonnais", "23" ] ] }
{ "S1": { "header": [ "Record ID", "Club" ], "rows": [ [ "R880", "SU Agen" ], [ "R262", "RC Toulonnais" ], [ "R134", "Biarritz Olympique" ], [ "R715", "Montpellier RC" ] ] }, ...
{ "S1": { "A": "For record R880, the club is SU Agen. For record R262, the club is RC Toulonnais. For record R134, the club is Biarritz Olympique. For record R715, the club is Montpellier RC.", "B": "| Record ID | Club |\n| --- | --- |\n| R880 | SU Agen |\n| R262 | RC Toulonnais |\n| R134 | Biarritz Olympique...
Among the supplied records, which club had the most losses?
en
[ "RC Toulonnais" ]
single_string
{ "AA": "For record R880, the club is SU Agen. For record R262, the club is RC Toulonnais. For record R134, the club is Biarritz Olympique. For record R715, the club is Montpellier RC.\n\nFor record R262, the number of losses is 23. For record R880, the number of losses is 11. For record R134, the number of losses is...
lang_table_10
argmin
{ "question_id": "nt-56", "question_file": "data/training.tsv", "original_question": "which member of the 500 club has the least amount of home runs?", "original_answers": [ "Eddie Murray" ], "table_name": "csv/203-csv/611.csv", "normalized_table_url": "https://raw.githubusercontent.com/ppasupat/WikiT...
{ "name": "csv/203-csv/611.tsv", "header": [ "Player", "HR" ], "rows": [ [ "Barry Bonds", "762" ], [ "Hank Aaron", "755" ], [ "Ted Williams", "521" ], [ "Gary Sheffield", "509" ], [ "Eddie Murray", "504" ...
{ "S1": { "header": [ "Record ID", "Player" ], "rows": [ [ "R268", "Ted Williams" ], [ "R389", "Gary Sheffield" ], [ "R942", "Hank Aaron" ], [ "R882", "Barry Bonds" ], [ ...
{ "S1": { "A": "For record R268, the player is Ted Williams. For record R389, the player is Gary Sheffield. For record R942, the player is Hank Aaron. For record R882, the player is Barry Bonds. For record R763, the player is Eddie Murray.", "B": "| Record ID | Player |\n| --- | --- |\n| R268 | Ted Williams |...
Among the supplied members of the 500 home run club, who has the fewest home runs?
en
[ "Eddie Murray" ]
single_string
{ "AA": "For record R268, the player is Ted Williams. For record R389, the player is Gary Sheffield. For record R942, the player is Hank Aaron. For record R882, the player is Barry Bonds. For record R763, the player is Eddie Murray.\n\nFor record R763, the home run count is 504. For record R389, the home run count is...

Prose–table representation transitions

Ten adapted WikiTableQuestions examples for PoC1 textual-representation experiments. Each example has four matched prompts (40 prompts total), using English prose (A) and Markdown tables (B). All inputs are text; there are no images.

Conditions and structure

Condition letters specify the representation of segments S1 and S2, respectively:

Condition S1 S2
AA A A
AB A B
BA B A
BB B B

Each question requires information from both segments. Meanings, identifiers, and the English question remain fixed across conditions. The fixed English prose question can introduce an additional boundary after S2 in AB and BB.

lang_table.json preserves the original project file, including top-level setup metadata and an examples list. The dataset configuration exposes that list as the test split. This split name denotes evaluation examples; no train/test sampling or held-out benchmark construction is implied.

Each example includes id, task, segments, question, question_language, and inputs. Gold answers are in expected_answers, with comparison rules in answer_comparison. Additional source, table, and segment_facts fields preserve provenance and auditing information.

Loading

from datasets import load_dataset

examples = load_dataset("talzoomanzoo/lang_table", split="test")
prompt = examples[0]["inputs"]["AB"]

To preserve the complete JSON document and its setup metadata:

import json
from huggingface_hub import hf_hub_download

path = hf_hub_download(
    repo_id="talzoomanzoo/lang_table",
    filename="lang_table.json",
    repo_type="dataset",
)
with open(path, encoding="utf-8") as f:
    document = json.load(f)
prompt = document["examples"][0]["inputs"]["AB"]

Send only the selected inputs value to the model. Gold answers and other metadata are for evaluation and auditing, and must not be included in the prompt.

Intended use and limitations

These small examples support setup validation for inference, token-position activation capture, representation comparisons, and probing experiments. They are not a representative benchmark. Establish model-specific accuracy in AA and BB before interpreting transition effects. Keep all four conditions of an example together when constructing any later evaluation partitions.

Real-world facts may be recalled without using the supplied segments. S1 maps opaque record IDs to entities or years; S2 maps IDs to attributes. Record IDs are independent of rank, year, and answer. Row orders are independently shuffled between segments, then held fixed across conditions.

Source, modifications, and license

Adapted from WikiTableQuestions by Panupong Pasupat and Percy Liang, introduced in Compositional Semantic Parsing on Semi-Structured Tables (2015). The upstream repository distributes the dataset under CC BY-SA 4.0; this adaptation retains that license.

Changes include selecting rows and task-relevant columns, rewording questions, introducing opaque record IDs, splitting facts into two segments, shuffling their row orders, and rendering matched prose and Markdown tables. Nonbreaking spaces are normalized to ordinary spaces. Original question IDs, questions, answers, table paths, source URLs, and one-based row numbers remain in each example's provenance. These are adapted examples, not an unmodified WikiTableQuestions evaluation split.

The design follows the project's PoC1 matched-input construction described in tue-meeting.pdf; that document is not included in this repository.

Downloads last month
-

Collection including talzoomanzoo/lang_table