level stringclasses 1
value | type stringlengths 11 95 | data_source stringclasses 1
value | prompt listlengths 1 1 | ability stringclasses 1
value | reward_model dict | extra_info dict |
|---|---|---|---|---|---|---|
DABench | dabstep_payments | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which merchant category had the highest payment refusal rate in 2023 among categories with at least 10,000 transactions? Return the MCC code and category description."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 0,
"dataset": "dabstep_payments",
"query_id": 1,
"query": "Which merchant category had the highest payment refusal rate in 2023 among categories with at least 10,000 transactions? Return the MCC code and category description.",
"prompt": "Which merchant category had the highest ... |
DABench | dabstep_payments | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "In 2023, which merchant had the most payments routed through an acquiring country that is not among its configured acquirer countries? Return the merchant name, acquiring country code, and number of such payments."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 1,
"dataset": "dabstep_payments",
"query_id": 2,
"query": "In 2023, which merchant had the most payments routed through an acquiring country that is not among its configured acquirer countries? Return the merchant name, acquiring country code, and number of such payments.",
"pro... |
DABench | dabstep_payments | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which merchant had the largest fraudulent cross-border transaction volume in 2023? Return the merchant name and the volume rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 2,
"dataset": "dabstep_payments",
"query_id": 3,
"query": "Which merchant had the largest fraudulent cross-border transaction volume in 2023? Return the merchant name and the volume rounded to two decimals.",
"prompt": "Which merchant had the largest fraudulent cross-border tran... |
DABench | dabstep_payments | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For Crossfit_Hanna, if every 2023 payment were priced as though its MCC had been 5411 for the entire year, what would be the increase in total processing fees compared with its actual MCC? Return the merchant name and the increase in euros rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 3,
"dataset": "dabstep_payments",
"query_id": 4,
"query": "For Crossfit_Hanna, if every 2023 payment were priced as though its MCC had been 5411 for the entire year, what would be the increase in total processing fees compared with its actual MCC? Return the merchant name and the ... |
DABench | dabstep_payments | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For Crossfit_Hanna in July 2023, which fee rule IDs apply to its transactions? Return all matching fee IDs in ascending order."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 4,
"dataset": "dabstep_payments",
"query_id": 5,
"query": "For Crossfit_Hanna in July 2023, which fee rule IDs apply to its transactions? Return all matching fee IDs in ascending order.",
"prompt": "For Crossfit_Hanna in July 2023, which fee rule IDs apply to its transactions? R... |
DABench | dabstep_payments | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "What total processing fee did Rafa_AI owe for March 2023? Return the amount in euros rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 5,
"dataset": "dabstep_payments",
"query_id": 6,
"query": "What total processing fee did Rafa_AI owe for March 2023? Return the amount in euros rounded to two decimals.",
"prompt": "What total processing fee did Rafa_AI owe for March 2023? Return the amount in euros rounded to t... |
DABench | olist_ecommerce | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Considering only orders with a status of delivered, which English product category generated the highest total product revenue (the sum of item prices, excluding freight)? Return the English category name and the total revenue rounded to two decimals. If two categories tie, ret... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 6,
"dataset": "olist_ecommerce",
"query_id": 1,
"query": "Considering only orders with a status of delivered, which English product category generated the highest total product revenue (the sum of item prices, excluding freight)? Return the English category name and the total reve... |
DABench | olist_ecommerce | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among delivered orders placed by customers located in the state of SP, what percentage of those orders were delivered to the customer after their estimated delivery date? Consider only orders with status delivered that have an actual customer delivery date. Return the percentag... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 7,
"dataset": "olist_ecommerce",
"query_id": 2,
"query": "Among delivered orders placed by customers located in the state of SP, what percentage of those orders were delivered to the customer after their estimated delivery date? Consider only orders with status delivered that have... |
DABench | olist_ecommerce | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Looking only at delivered order items, and considering English product categories that have at least 1000 such items, which category has the highest average review score? Each delivered order item takes the review score of its order, and an order's review score is the average o... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 8,
"dataset": "olist_ecommerce",
"query_id": 3,
"query": "Looking only at delivered order items, and considering English product categories that have at least 1000 such items, which category has the highest average review score? Each delivered order item takes the review score of ... |
DABench | olist_ecommerce | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Considering only items belonging to orders with a status of delivered, which seller state generated the highest total product revenue (the sum of item prices, excluding freight)? Return the two-letter seller state code and the total revenue rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 9,
"dataset": "olist_ecommerce",
"query_id": 4,
"query": "Considering only items belonging to orders with a status of delivered, which seller state generated the highest total product revenue (the sum of item prices, excluding freight)? Return the two-letter seller state code and ... |
DABench | olist_ecommerce | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among orders with a status of delivered, how many have at least one customer review whose comment message mentions a delay (the message contains the substring 'atrasad', case-insensitive, covering forms such as 'atrasado' and 'atrasada')? Return the count of distinct delivered ... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 10,
"dataset": "olist_ecommerce",
"query_id": 5,
"query": "Among orders with a status of delivered, how many have at least one customer review whose comment message mentions a delay (the message contains the substring 'atrasad', case-insensitive, covering forms such as 'atrasado' ... |
DABench | olist_ecommerce | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Looking only at delivered order items priced above zero, and considering English product categories that have at least 1000 such items, which category has the highest average freight-to-price ratio (the per-item freight value divided by its price, averaged over the items)? Retu... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 11,
"dataset": "olist_ecommerce",
"query_id": 6,
"query": "Looking only at delivered order items priced above zero, and considering English product categories that have at least 1000 such items, which category has the highest average freight-to-price ratio (the per-item freight va... |
DABench | pitchfork_reviews | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Considering only genres that have at least 500 scored reviews, which genre has the highest average review score? Return the genre and the average score rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 12,
"dataset": "pitchfork_reviews",
"query_id": 1,
"query": "Considering only genres that have at least 500 scored reviews, which genre has the highest average review score? Return the genre and the average score rounded to two decimals.",
"prompt": "Considering only genres that... |
DABench | pitchfork_reviews | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among albums that were named best new music, how many have a review whose full text mentions the word 'masterpiece' (case-insensitive)? Return the count."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 13,
"dataset": "pitchfork_reviews",
"query_id": 2,
"query": "Among albums that were named best new music, how many have a review whose full text mentions the word 'masterpiece' (case-insensitive)? Return the count.",
"prompt": "Among albums that were named best new music, how ma... |
DABench | pitchfork_reviews | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Looking only at reviews published in 2016, and considering only record labels with at least 20 such reviews, which label has the highest average review score? Return the label and the average score rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 14,
"dataset": "pitchfork_reviews",
"query_id": 3,
"query": "Looking only at reviews published in 2016, and considering only record labels with at least 20 such reviews, which label has the highest average review score? Return the label and the average score rounded to two decimal... |
DABench | pitchfork_reviews | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "A review is considered a reissue when the album's earliest listed release year is at least 20 years before the year the review was published. How many reviews are reissues by this definition? Return the count."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 15,
"dataset": "pitchfork_reviews",
"query_id": 4,
"query": "A review is considered a reissue when the album's earliest listed release year is at least 20 years before the year the review was published. How many reviews are reissues by this definition? Return the count.",
"promp... |
DABench | pitchfork_reviews | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among albums that were named best new music, what is the average character length of their full review text? Return the average rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 16,
"dataset": "pitchfork_reviews",
"query_id": 5,
"query": "Among albums that were named best new music, what is the average character length of their full review text? Return the average rounded to two decimals.",
"prompt": "Among albums that were named best new music, what is... |
DABench | pitchfork_reviews | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Assign each review to a release decade based on the album's earliest listed release year (for example, a 1993 release belongs to the 1990 decade). Considering only decades that have at least 300 scored reviews, which release decade has the highest average review score? Return t... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 17,
"dataset": "pitchfork_reviews",
"query_id": 6,
"query": "Assign each review to a release decade based on the album's earliest listed release year (for example, a 1993 release belongs to the 1990 decade). Considering only decades that have at least 300 scored reviews, which rel... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For each IPL season, identify the three highest run scorers and the three highest credited wicket takers, then pair them by rank: run scorer 1 with wicket taker 1, run scorer 2 with wicket taker 2, and run scorer 3 with wicket taker 3. Break ties by the smaller player ID. Retur... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 18,
"dataset": "spider2_IPL",
"query_id": 1,
"query": "For each IPL season, identify the three highest run scorers and the three highest credited wicket takers, then pair them by rank: run scorer 1 with wicket taker 1, run scorer 2 with wicket taker 2, and run scorer 3 with wicket... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "In the last match of each IPL season, which partnership had the highest batsman-run total? Return the season, match ID, innings, both player names and IDs, each player's partnership runs, and the total partnership runs. List player 1 as the partner with the higher individual sc... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 19,
"dataset": "spider2_IPL",
"query_id": 2,
"query": "In the last match of each IPL season, which partnership had the highest batsman-run total? Return the season, match ID, innings, both player names and IDs, each player's partnership runs, and the total partnership runs. List p... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Across overs that were the highest-scoring over within their respective match, which three bowlers conceded the largest single-over totals? Return the bowler name, player ID, match ID, innings, over, and runs conceded."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 20,
"dataset": "spider2_IPL",
"query_id": 3,
"query": "Across overs that were the highest-scoring over within their respective match, which three bowlers conceded the largest single-over totals? Return the bowler name, player ID, match ID, innings, over, and runs conceded.",
"pr... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among bowlers with at least 50 credited wickets, which bowler has the lowest bowling average per wicket? Return the bowler name, player ID, bowling average, runs conceded, and credited wickets."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 21,
"dataset": "spider2_IPL",
"query_id": 4,
"query": "Among bowlers with at least 50 credited wickets, which bowler has the lowest bowling average per wicket? Return the bowler name, player ID, bowling average, runs conceded, and credited wickets.",
"prompt": "Among bowlers wit... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among right-hand batters who scored more than 50 runs in at least one match, what is the average of their career IPL striker-run totals? Return the eligible player count and the average rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 22,
"dataset": "spider2_IPL",
"query_id": 5,
"query": "Among right-hand batters who scored more than 50 runs in at least one match, what is the average of their career IPL striker-run totals? Return the eligible player count and the average rounded to two decimals.",
"prompt": "... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which players scored at least 100 runs in a match while playing for the team that lost that match? Return all matching match IDs, player names, player IDs, team names, and runs."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 23,
"dataset": "spider2_IPL",
"query_id": 6,
"query": "Which players scored at least 100 runs in a match while playing for the team that lost that match? Return all matching match IDs, player names, player IDs, team names, and runs.",
"prompt": "Which players scored at least 100... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "In season 5, which five players had the highest average runs per match among players with at least five batting matches? Return player names, player IDs, matches, total runs, and averages rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 24,
"dataset": "spider2_IPL",
"query_id": 7,
"query": "In season 5, which five players had the highest average runs per match among players with at least five batting matches? Return player names, player IDs, matches, total runs, and averages rounded to two decimals.",
"prompt":... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For countries with at least four right-hand batters who each batted in at least ten matches, which five countries have the highest average of those players' per-match batting averages? Return the country, eligible player count, and country average rounded to two decimals."
}
... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 25,
"dataset": "spider2_IPL",
"query_id": 8,
"query": "For countries with at least four right-hand batters who each batted in at least ten matches, which five countries have the highest average of those players' per-match batting averages? Return the country, eligible player count... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For every result match, combine batsman runs and extras for each over, take the single highest over total in that match, and return the number of matches and the average of those match-level maxima rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 26,
"dataset": "spider2_IPL",
"query_id": 9,
"query": "For every result match, combine batsman runs and extras for each over, take the single highest over total in that match, and return the number of matches and the average of those match-level maxima rounded to two decimals.",
... |
DABench | spider2_IPL | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "In season 7, which captain scored the largest share of his team's batsman runs in a match his team lost? Return the captain name, player ID, match ID, team name, captain runs, team runs, and share rounded to four decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 27,
"dataset": "spider2_IPL",
"query_id": 10,
"query": "In season 7, which captain scored the largest share of his team's batsman runs in a match his team lost? Return the captain name, player ID, match ID, team name, captain runs, team runs, and share rounded to four decimals.",
... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "What is the distance of the longest route where Abakan is either the departure or destination city? Return the two endpoint cities and the distance in kilometers rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 28,
"dataset": "spider2_airlines",
"query_id": 1,
"query": "What is the distance of the longest route where Abakan is either the departure or destination city? Return the two endpoint cities and the distance in kilometers rounded to two decimals.",
"prompt": "What is the distanc... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Distribute all unique city pairs into distance ranges 0-999, 1000-1999, 2000-2999, 3000-3999, 4000-4999, 5000-5999, and 6000+, based on the average Haversine distance of all flights between the two cities. Which range has the fewest unique city pairs? Return the range and the p... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 29,
"dataset": "spider2_airlines",
"query_id": 2,
"query": "Distribute all unique city pairs into distance ranges 0-999, 1000-1999, 2000-2999, 3000-3999, 4000-4999, 5000-5999, and 6000+, based on the average Haversine distance of all flights between the two cities. Which range has... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among aircraft models with at least 1000 fare coupons on arrived flights longer than 3000 km, which model has the highest average passenger fare per kilometer? Return the model and the average fare per kilometer rounded to four decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 30,
"dataset": "spider2_airlines",
"query_id": 3,
"query": "Among aircraft models with at least 1000 fare coupons on arrived flights longer than 3000 km, which model has the highest average passenger fare per kilometer? Return the model and the average fare per kilometer rounded t... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For arrived flights whose departure and arrival airports are in different timezones, which departure city generated the most Business fare revenue? Return the departure city and total revenue rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 31,
"dataset": "spider2_airlines",
"query_id": 4,
"query": "For arrived flights whose departure and arrival airports are in different timezones, which departure city generated the most Business fare revenue? Return the departure city and total revenue rounded to two decimals.",
... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among arrived flights operated by aircraft with a maximum range under 3000 km, which city pair has the longest average Haversine route distance? Consider only city pairs with at least five flights. Return the two cities and the average distance rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 32,
"dataset": "spider2_airlines",
"query_id": 5,
"query": "Among arrived flights operated by aircraft with a maximum range under 3000 km, which city pair has the longest average Haversine route distance? Consider only city pairs with at least five flights. Return the two cities a... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which city pair has the highest average Comfort fare per kilometer among city pairs with at least 50 Comfort tickets on arrived flights? Return the two cities and the average fare per kilometer rounded to four decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 33,
"dataset": "spider2_airlines",
"query_id": 6,
"query": "Which city pair has the highest average Comfort fare per kilometer among city pairs with at least 50 Comfort tickets on arrived flights? Return the two cities and the average fare per kilometer rounded to four decimals.",... |
DABench | spider2_airlines | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For bookings made from August 1 through August 7, 2017, which departure city generated the most arrived-flight fare revenue? Return the departure city and total revenue rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 34,
"dataset": "spider2_airlines",
"query_id": 7,
"query": "For bookings made from August 1 through August 7, 2017, which departure city generated the most arrived-flight fare revenue? Return the departure city and total revenue rounded to two decimals.",
"prompt": "For bookings... |
DABench | spider2_bowling_league | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which bowlers won games with handicap scores of 190 or less at all three venues: Thunderbird Lanes, Totem Lanes, and Bolero Lanes? Return only the qualifying game records at those three venues with bowler roster ID, first name, last name, match score ID, game code, handicap sco... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 35,
"dataset": "spider2_bowling_league",
"query_id": 1,
"query": "Which bowlers won games with handicap scores of 190 or less at all three venues: Thunderbird Lanes, Totem Lanes, and Bolero Lanes? Return only the qualifying game records at those three venues with bowler roster ID,... |
DABench | spider2_bowling_league | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For each tournament venue, which team recorded the most match-game wins? Break ties alphabetically by team name to choose the venue winner. Return venue, team roster ID, team name, and win count. Row order is not important."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 36,
"dataset": "spider2_bowling_league",
"query_id": 2,
"query": "For each tournament venue, which team recorded the most match-game wins? Break ties alphabetically by team name to choose the venue winner. Return venue, team roster ID, team name, and win count. Row order is not im... |
DABench | spider2_bowling_league | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among registered cities with at least three ZIP-confirmed bowlers who have game score records, which cities have a city-level average of individual bowler season raw-score averages above the roster current average for those same scoring bowlers? Return city, confirmed scoring-b... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 37,
"dataset": "spider2_bowling_league",
"query_id": 3,
"query": "Among registered cities with at least three ZIP-confirmed bowlers who have game score records, which cities have a city-level average of individual bowler season raw-score averages above the roster current average f... |
DABench | spider2_bowling_league | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among teams that recorded match-game wins at every tournament venue, identify the top five teams by how many low-handicap games their captain personally won, where low-handicap means a winning handicap score of 190 or less. Use total team match-game wins descending, then team n... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 38,
"dataset": "spider2_bowling_league",
"query_id": 4,
"query": "Among teams that recorded match-game wins at every tournament venue, identify the top five teams by how many low-handicap games their captain personally won, where low-handicap means a winning handicap score of 190 ... |
DABench | spider2_bowling_league | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which five bowlers exceeded their roster current average at the most tournament venues, where a venue counts only if the bowler's average raw score at that venue is greater than the current average? Use largest positive venue gap descending, then last name and first name, only ... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 39,
"dataset": "spider2_bowling_league",
"query_id": 5,
"query": "Which five bowlers exceeded their roster current average at the most tournament venues, where a venue counts only if the bowler's average raw score at that venue is greater than the current average? Use largest posi... |
DABench | spider2_bowling_league | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For each tournament month, which team had the most match-game wins? Break ties alphabetically by team name. Return month, team ID, team name, win count, and number of tournaments held that month. Row order is not important."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 40,
"dataset": "spider2_bowling_league",
"query_id": 6,
"query": "For each tournament month, which team had the most match-game wins? Break ties alphabetically by team name. Return month, team ID, team name, win count, and number of tournaments held that month. Row order is not im... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among collisions with an at-fault driver whose sobriety notes indicate they had been drinking, which primary violation category appears most often? Return the category and the number of distinct collisions."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 41,
"dataset": "spider2_california_traffic",
"query_id": 1,
"query": "Among collisions with an at-fault driver whose sobriety notes indicate they had been drinking, which primary violation category appears most often? Return the category and the number of distinct collisions.",
... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among counties with at least 20 at-fault driver records where age is available, which county has the highest share of those drivers younger than 25? Return the county and the percentage rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 42,
"dataset": "spider2_california_traffic",
"query_id": 2,
"query": "Among counties with at least 20 at-fault driver records where age is available, which county has the highest share of those drivers younger than 25? Return the county and the percentage rounded to two decimals."... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For hit-and-run collisions with a participant record saying no proof of insurance was obtained, which severity level is most common? Return the severity and the number of distinct collisions."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 43,
"dataset": "spider2_california_traffic",
"query_id": 3,
"query": "For hit-and-run collisions with a participant record saying no proof of insurance was obtained, which severity level is most common? Return the severity and the number of distinct collisions.",
"prompt": "For ... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For non-daylight broadside collisions, which collision year has the highest average age among injured victims? Consider only years with at least three injured victim records. Return the year and the average age rounded to two decimals."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 44,
"dataset": "spider2_california_traffic",
"query_id": 4,
"query": "For non-daylight broadside collisions, which collision year has the highest average age among injured victims? Consider only years with at least three injured victim records. Return the year and the average age ... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which vehicle make appears most often as the at-fault driver make in fatal or severe-injury collisions? Return the make and the number of distinct collisions."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 45,
"dataset": "spider2_california_traffic",
"query_id": 5,
"query": "Which vehicle make appears most often as the at-fault driver make in fatal or severe-injury collisions? Return the make and the number of distinct collisions.",
"prompt": "Which vehicle make appears most often... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "In motorcycle collisions, among victims whose equipment notes mention that a helmet was used, which injury degree is most common? Return the injury degree and the victim count."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 46,
"dataset": "spider2_california_traffic",
"query_id": 6,
"query": "In motorcycle collisions, among victims whose equipment notes mention that a helmet was used, which injury degree is most common? Return the injury degree and the victim count.",
"prompt": "In motorcycle colli... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "For collisions involving victims whose equipment notes mention child restraint in vehicle used, which collision type is most common? Return the collision type and the number of distinct collisions."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 47,
"dataset": "spider2_california_traffic",
"query_id": 7,
"query": "For collisions involving victims whose equipment notes mention child restraint in vehicle used, which collision type is most common? Return the collision type and the number of distinct collisions.",
"prompt":... |
DABench | spider2_california_traffic | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among collision hours, which hour has the most injured victim records, and how many of those injured victims were younger than 18? Return the hour, the injured victim count, and the under-18 count."
}
] | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 48,
"dataset": "spider2_california_traffic",
"query_id": 8,
"query": "Among collision hours, which hour has the most injured victim records, and how many of those injured victims were younger than 18? Return the hour, the injured victim count, and the under-18 count.",
"prompt":... |
DABench | spider2_delivery_center | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among hub cities with at least 1,000 finished FOOD orders with paid payments and completed deliveries in both February and March, which city had the largest increase in average paid basket from February to March? Return the city, February order count, February average paid bask... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 49,
"dataset": "spider2_delivery_center",
"query_id": 1,
"query": "Among hub cities with at least 1,000 finished FOOD orders with paid payments and completed deliveries in both February and March, which city had the largest increase in average paid basket from February to March? R... |
DABench | spider2_delivery_center | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which store segment and channel type combination had a higher cancellation rate in April than January, while its average cycle time for finished orders decreased over the same months? Consider only combinations with at least 500 orders in both months. Return segment, channel ty... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 50,
"dataset": "spider2_delivery_center",
"query_id": 2,
"query": "Which store segment and channel type combination had a higher cancellation rate in April than January, while its average cycle time for finished orders decreased over the same months? Consider only combinations wit... |
DABench | spider2_delivery_center | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Among stores with at least 150 finished orders in every month from January through April, which five stores had the largest absolute paid-minus-order gap per paid-backed finished order? For the reconciliation totals, include only finished orders that have paid payment records. ... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 51,
"dataset": "spider2_delivery_center",
"query_id": 3,
"query": "Among stores with at least 150 finished orders in every month from January through April, which five stores had the largest absolute paid-minus-order gap per paid-backed finished order? For the reconciliation total... |
DABench | spider2_delivery_center | dab_sandbox_hard200_plus_yuchen80 | [
{
"role": "user",
"content": "Which ten drivers with at least 50 completed deliveries in each month had an OWN CHANNEL delivery share that increased every month from January through April? Return driver token, modal, type, January share, February share, March share, April share, and total completed deliveri... | data_agent_bench | {
"style": "rule",
"ground_truth": ""
} | {
"split": "train",
"index": 52,
"dataset": "spider2_delivery_center",
"query_id": 4,
"query": "Which ten drivers with at least 50 completed deliveries in each month had an OWN CHANNEL delivery share that increased every month from January through April? Return driver token, modal, type, January share, Februa... |
End of preview. Expand in Data Studio
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
DAB sandbox hard168 + Yuchen80 VERL dataset
Public DAB sandbox / VERL delta package.
Rows: 248 train / 248 test.
Sources:
- hard200 Qwen3.5 selected training-hard: 168 rows.
- Yuchen DAB task expansion all80: 80 rows.
Files:
train.parquet,test.parquet: VERL-readable rows.task_artifacts/: DAB task directories with validators and only the query_dataset assets missing from the current sandbox base.sandbox_task_manifest.json,task_artifacts_manifest.json,task_list.json.scripts/materialize_dabench_assets.py: materialize task artifacts into a DataAgentBench root.
Package mode:
- This is not a fully self-contained copy of every DB.
- It assumes the target sandbox/DataAgentBench already has the standard DAB assets and the Yuchen expansion DBs that are already present there.
- It only ships the Yuchen expansion
query_datasetfolders that were missing from the current sandbox base:query_pitchfork_reviewsquery_spider2_airlinesquery_spider2_bowling_leaguequery_spider2_california_trafficquery_spider2_delivery_centerquery_spider2_imdb_moviesquery_spider2_stacking
Materialize into an existing sandbox/DataAgentBench root:
python scripts/materialize_dabench_assets.py --package-dir /path/to/dab-hard168-yuchen80 --bench-root /mnt/docker-data/workspace/DataAgentBench --canonical-bench-root /mnt/docker-data/workspace/DataAgentBench --overwrite
The script merges task files without deleting existing query_dataset directories. Missing query_dataset folders are either copied from this package or linked from the existing canonical sandbox assets.
Then mount /mnt/docker-data/workspace/DataAgentBench into the sandbox as /workspace/DataAgentBench and point VERL at train.parquet.
- Downloads last month
- 352