Wide aggregation into a table
Collect many entities and attributes across sites into a complete table.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- Chrome · spreadsheets
- Class GUI-need label
CL· CLI-sufficient
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R04.01.01.02-T01 EU statutory minimum wages table
Original instruction
Build minwage.csv with one row per current EU member state: columns country, has_statutory_minimum (yes/no), monthly_minimum_eur. For states without a national statutory minimum, set has_statutory_minimum=no and leave the amount blank. Use the latest published figures.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals the 27 member states (recall+precision 1.0); has_statutory flags correct; amounts within 2% of reference; no-statutory states correctly blank. Graded against the reference CSV.
Proposed partial credit
30% all 27 states; 30% statutory flags; 40% amounts within tolerance
R04.01.01.02-T02 African capitals with population
Original instruction
Build africapitals.csv with one row per sovereign African country: columns country, capital, capital_population. Use the most recent available population for the capital city. Mark any capital whose population figure is unavailable as blank.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizon- Proposed difficulty
- 4 / 5
Proposed verifier
Row set equals the 54 sovereign states (recall+precision 1.0); capitals correct; populations within 5% of reference or correctly blank. Graded against the reference CSV.
Proposed partial credit
30% all countries; 30% correct capitals; 40% populations within tolerance
R04.01.01.02-T03 World Cup finals table
Original instruction
Build wcfinals.csv with one row per FIFA men's World Cup final: columns year, host_country, winner, runner_up, final_score. Include every edition through the latest completed tournament.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
Row set equals every edition (recall+precision 1.0); winner/runner-up/score cells match reference exactly. Graded against the reference CSV.
Proposed partial credit
40% all editions; 30% winners/runners; 30% scores
R04.01.01.02-T04 Periodic table attributes
Original instruction
Build elements.csv with one row per chemical element (1-118): columns atomic_number, symbol, name, group, period. Leave group blank for the f-block lanthanides/actinides where no main group applies.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
118 rows with correct symbol/name/period for each atomic number and correct group (blank for f-block). Graded against the reference CSV (exact).
Proposed partial credit
30% all 118 rows; 40% symbol/name/period; 30% group handling
R04.01.01.02-T05 US states facts table
Original instruction
Build states.csv with one row per US state: columns state, postal_abbreviation, capital, year_admitted. Fifty rows, no territories.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
CL· CLI-sufficient- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 2 / 5
Proposed verifier
50 rows (no territories); abbreviation, capital, and admission year correct for each. Graded against the reference CSV (exact).
Proposed partial credit
30% all 50 states; 40% capitals+abbreviations; 30% admission years
R04.01.01.02-T06 Nobel Physics laureates in a decade
Original instruction
Build nobel_physics.csv with one row per Nobel Prize in Physics awarded in a specified decade: columns year, laureate, country_affiliation, citation_topic. Split shared prizes into one row per laureate.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals all laureates in the decade (recall+precision 1.0, shared prizes split); country and topic cells match reference. Graded against the reference CSV.
Proposed partial credit
40% all laureates (splits); 30% countries; 30% topics
R04.01.01.02-T07 Metro line stations table
Original instruction
Build line_stations.csv with one row per station on a named metro line, in order: columns position, station_name, opening_year, interchange_lines. List interchange lines as a semicolon-separated set (blank if none).
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals every station in correct order (recall+precision 1.0); opening years and interchange sets match reference. Graded against the reference CSV.
Proposed partial credit
40% all stations in order; 30% opening years; 30% interchanges
R04.01.01.02-T08 UNESCO sites in a country
Original instruction
Build unesco.csv with one row per UNESCO World Heritage Site in a specified country: columns site_name, year_inscribed, category (cultural/natural/mixed). Include all inscribed sites through the latest session.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals all inscribed sites (recall+precision 1.0); year and category correct. Graded against the reference CSV.
Proposed partial credit
40% all sites; 30% years; 30% categories
R04.01.01.02-T09 Summer Olympics editions table
Original instruction
Build olympics.csv with one row per Summer Olympic Games: columns year, host_city, host_country, nations_participating, events. Include every held edition (exclude cancelled ones, noting none).
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals every held edition (recall+precision 1.0); city/country exact; nations and events within reference values. Graded against the reference CSV.
Proposed partial credit
40% all editions+cities; 30% countries; 30% nations/events counts
R04.01.01.02-T10 Director filmography table
Original instruction
Build filmography.csv with one row per feature film directed by a named director: columns title, year, runtime_min, distributor. Include only theatrical features (exclude shorts and TV), noting exclusions are intentional.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals the director's theatrical features (recall+precision 1.0); year and runtime correct; distributor matches reference. Graded against the reference CSV.
Proposed partial credit
40% all features; 30% years; 30% runtime/distributor
R04.01.01.02-T11 Counties of a state table
Original instruction
Build counties.csv with one row per county in a specified US state: columns county_name, county_seat, population_latest_census. Include every county (or equivalent).
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizon- Proposed difficulty
- 4 / 5
Proposed verifier
Row set equals every county (recall+precision 1.0); seats correct; populations match the latest census within 1%. Graded against the reference CSV.
Proposed partial credit
40% all counties; 30% seats; 30% populations
R04.01.01.02-T12 Grand Slam champions over a span
Original instruction
Build slam_champions.csv with one row per Grand Slam singles title in a specified multi-year span: columns year, tournament, draw (men/women), champion, runner_up. One row per final.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizon- Proposed difficulty
- 4 / 5
Proposed verifier
Row set equals all finals in the span (recall+precision 1.0); champion/runner-up cells match reference. Graded against the reference CSV.
Proposed partial credit
40% all finals; 30% champions; 30% runners-up
R04.01.01.02-T13 Planets and major moons table
Original instruction
Build bodies.csv with one row per planet and per major moon (diameter > 400 km) of the solar system: columns body_name, type (planet/moon), parent_body, mean_radius_km, discovery_year. Planets have blank discovery_year where applicable.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals the planets + qualifying moons (recall+precision 1.0 against the reference threshold list); radii within 2%; discovery years correct or blank. Graded against the reference CSV.
Proposed partial credit
40% all bodies; 30% radii; 30% discovery years
R04.01.01.02-T14 Constituencies and winners table
Original instruction
Build constituencies.csv with one row per electoral constituency in a specified region for a given election: columns constituency, winning_party, winning_candidate, margin_votes. Public election results only.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizon- Proposed difficulty
- 4 / 5
Proposed verifier
Row set equals every constituency (recall+precision 1.0); winning party/candidate correct; margins within 1%. Graded against the reference CSV.
Proposed partial credit
40% all constituencies; 30% winning party; 30% candidate+margin
R04.01.01.02-T15 Band discography table
Original instruction
Build discography.csv with one row per studio album by a named band: columns title, release_year, label, peak_chart_position. Studio albums only (exclude live/compilations), exclusions intentional.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals the studio albums (recall+precision 1.0); year/label correct; chart peak matches reference or blank where uncharted. Graded against the reference CSV.
Proposed partial credit
40% all studio albums; 30% year/label; 30% chart peaks
R04.01.01.02-T16 Airports of a country table
Original instruction
Build airports.csv with one row per airport with scheduled passenger service in a specified country: columns airport_name, iata_code, city_served, type (international/domestic). Include every such airport.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizon- Proposed difficulty
- 4 / 5
Proposed verifier
Row set equals all qualifying airports (recall >= 0.95, precision 1.0); IATA codes and city served correct. Graded against the reference CSV.
Proposed partial credit
40% airport recall; 30% IATA codes; 30% city/type
R04.01.01.02-T17 Eclipses in a date range table
Original instruction
Build eclipses.csv with one row per solar eclipse in a specified multi-year range: columns date, type (total/annular/partial/hybrid), magnitude, main_region_of_visibility. Include every solar eclipse in the range.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 4 / 5
Proposed verifier
Row set equals all solar eclipses in the range (recall+precision 1.0); type and magnitude match reference; region correct. Graded against the reference CSV.
Proposed partial credit
40% all eclipses; 30% types; 30% magnitude/region
R04.01.01.02-T18 Legislative committee roster table
Original instruction
Build committee.csv with one row per member of a specified public legislative committee: columns member_name, party, district_or_state, role (chair/ranking/member). Use the current published roster.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & forms- Work product
WP02· Records/table- Conditions
- None declared
- Proposed difficulty
- 3 / 5
Proposed verifier
Row set equals the current roster (recall+precision 1.0); party, district, and role correct. Graded against the reference CSV.
Proposed partial credit
40% all members; 30% party/district; 30% roles
R04.01.01.02-T19 Full season results table
Original instruction
Build season.csv with one row per match of a specified sports league regular season: columns matchday, date, home_team, away_team, home_score, away_score. Include every scheduled match (hundreds of rows).
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
HY· Hybrid- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Row set equals every scheduled match (recall >= 0.98, precision 1.0); scores and teams match reference per match. Graded against the reference CSV.
Proposed partial credit
40% match recall; 30% correct fixtures; 30% scores
R04.01.01.02-T20 Multi-site attribute join table
Original instruction
Build companies.csv for a named stock index: one row per constituent with columns company, ticker, sector, headquarters_city, founded_year. Each attribute may require a different source (index list, company profile, encyclopedia). Mark any unavailable cell blank.
- Platforms
web· Any OS (web)- Applications
- Chrome · spreadsheets
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
CL· CLI-sufficient- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollclick· Point & clicktype· Text & formsmulti· Cross-app- Work product
WP02· Records/table- Conditions
long· Long-horizonmultiapp· Multi-app- Proposed difficulty
- 5 / 5
Proposed verifier
Row set equals the index constituents (recall+precision 1.0); ticker/sector/HQ/founded match reference or correctly blank. Graded against the reference CSV.
Proposed partial credit
30% all constituents; 30% tickers/sectors; 40% HQ + founded year
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.