datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
dbbench-spider-3500
DBBench-Spider-3500
AgentBench DBBench 評価ハーネスと完全互換のフォーマットで生成した SFT 訓練データセット。
Spider データセット (Yale NLP) の 3,500 問を GPT-OSS-120B (Groq) に解かせ、正解したトラジェクトリ 1,697 件 を収録。
混合利用を想定: 本データセットは mark-22/dbbench_cleaned_for_agentbench（1,200 件）と混合し、合計 2,897 件 の SFT データとして使用することを想定しています。
Dataset Summary
Metric
Value
Total trajectories
1,697
Difficulty: Medium
1,406
Difficulty: Hard
291
Avg messages per item
13.2
Unique databases (db_id)
159
Source questions3… See the full description on the dataset page: https://huggingface.co/datasets/mark-22/dbbench-spider-3500.dbbench_sft_dataset_react
DBBench SFT Dataset (ReAct Format — AgentBench Compatible)
Overview
Synthetic SFT dataset for DBBench (AgentBench, ICLR 2024).
All tables, data, and queries are independently generated to avoid test data leakage.
Format
ReAct text format matching the AgentBench DBBench evaluation protocol:
[user] System prompt (Action: Operation / Action: Answer instructions)
[agent] Ok.
[user] Question + table name + column headers
[agent] Thinking + Action: Operation +… See the full description on the dataset page: https://huggingface.co/datasets/u-10bei/dbbench_sft_dataset_react.dbbench_sft_dataset_react_v2
DBBench SFT Dataset (ReAct Format — AgentBench Compatible)
Overview
Synthetic SFT dataset for DBBench (AgentBench, ICLR 2024).
All tables, data, and queries are independently generated to avoid test data leakage.
Format
ReAct text format matching the AgentBench DBBench evaluation protocol:
[user] System prompt (Action: Operation / Action: Answer instructions)
[agent] Ok.
[user] Question + table name + column headers
[agent] Thinking + Action: Operation +… See the full description on the dataset page: https://huggingface.co/datasets/u-10bei/dbbench_sft_dataset_react_v2.dbbench_and_alfworld_sft_dataset_v4
DBBench + ALFWorld SFT Dataset (Merged)
Overview
This dataset is a simple concatenation (merge) of the following two synthetic SFT datasets:
ALFWorld Trajectory Dataset: moroqq/sft_alfworld_trajectory_dataset_v2
https://huggingface.co/datasets/moroqq/sft_alfworld_trajectory_dataset_v2
DBBench SFT Dataset (ReAct Format): u-10bei/dbbench_sft_dataset_react_v4
https://huggingface.co/datasets/u-10bei/dbbench_sft_dataset_react_v4
The goal is to provide a single dataset… See the full description on the dataset page: https://huggingface.co/datasets/moroqq/dbbench_and_alfworld_sft_dataset_v4.dbbench_sft_dataset_react_v4
DBBench SFT Dataset (ReAct Format — AgentBench Compatible)
Overview
Synthetic SFT dataset for DBBench (AgentBench, ICLR 2024).
All tables, data, and queries are independently generated to avoid test data leakage.
Format
ReAct text format matching the AgentBench DBBench evaluation protocol:
[user] System prompt (Action: Operation / Action: Answer instructions)
[agent] Ok.
[user] Question + table name + column headers
[agent] Thinking + Action: Operation +… See the full description on the dataset page: https://huggingface.co/datasets/u-10bei/dbbench_sft_dataset_react_v4.dbbench_sft_dataset200
DBBench SFT Dataset 200 (Robustness-Heavy Subset)
This dataset is a 200-sample subset of u-10bei/dbbench_sft_dataset_react_v4.
Motivation
You want to mix a small amount of DBBench data into ALFWorld SFT to reduce DBBench regression.
This subset is biased toward retry / error-recovery trajectories to teach robustness.
Selection Policy (Summary)
Total: 200
Seed: 0
Max samples per metadata.table_name: 2
Target allocation (approx.):
{"core_insert": 40… See the full description on the dataset page: https://huggingface.co/datasets/moroqq/dbbench_sft_dataset200.dbbench_and_alfworld_sft_dataset
DBBench + ALFWorld SFT Dataset (Merged)
Overview
This dataset is a simple concatenation (merge) of the following two synthetic SFT datasets:
ALFWorld Trajectory Dataset: u-10bei/sft_alfworld_trajectory_dataset_v5
https://huggingface.co/datasets/u-10bei/sft_alfworld_trajectory_dataset_v5
DBBench SFT Dataset (ReAct Format): u-10bei/dbbench_sft_dataset_react_v4
https://huggingface.co/datasets/u-10bei/dbbench_sft_dataset_react_v4
The goal is to provide a single dataset… See the full description on the dataset page: https://huggingface.co/datasets/moroqq/dbbench_and_alfworld_sft_dataset.dbbench_cleaned_for_agentbench
DBBench Cleaned for AgentBench
u-10bei/dbbench_sft_dataset_react_v4（1,200 件）に対してクレンジング処理を施したデータセット。
AgentBench DBBench 評価用の SFT 訓練データとしてそのまま使用可能。
混合利用を想定: 本データセットは mark-22/dbbench-spider-3500（1,697 件）と混合し、合計 2,897 件 の SFT データとして使用することを想定しています。
Dataset Summary
Metric
Value
Total rows
1,200
Source
u-10bei/dbbench_sft_dataset_react_v4
Avg messages per item
6.7
Items with Final Answer1,200 / 1,200 (100%)
Columns
id, messages, metadata… See the full description on the dataset page: https://huggingface.co/datasets/mark-22/dbbench_cleaned_for_agentbench.dbbench_sft_dataset_org_v3
dbbench_sft_dataset_org_v3
A DBBench-style SFT dataset combining 295 stratified examples from dbbench_u-10bei_sft_dataset_modified_v1 (100 from INSERT/INSERT_error_recovery + 195 from other types) and 5 handcrafted examples from dbbench_sft_dataset_org_v1.
Overview
Total: 300 conversations
100 randomly sampled from INSERT and INSERT_error_recovery types in modified_v1
195 randomly sampled from other types (UPDATE, aggregation, comparison, etc.) in modified_v1
5… See the full description on the dataset page: https://huggingface.co/datasets/ShogoMu/dbbench_sft_dataset_org_v3.dbbench_and_alfworld_sft_dataset_v3
DBBench + ALFWorld SFT Dataset (Merged)
Overview
This dataset is a simple concatenation (merge) of the following two synthetic SFT datasets:
ALFWorld Trajectory Dataset: moroqq/sft_alfworld_trajectory_dataset_v2
https://huggingface.co/datasets/moroqq/sft_alfworld_trajectory_dataset_v2
DBBench SFT Dataset (ReAct Format): moroqq/dbbench_sft_dataset200
https://huggingface.co/datasets/moroqq/dbbench_sft_dataset200
The goal is to provide a single dataset repo that… See the full description on the dataset page: https://huggingface.co/datasets/moroqq/dbbench_and_alfworld_sft_dataset_v3.dbbench_sft_dataset_react_v3v4_weaktypes
DBBench SFT Trajectories (Weak Types from v3–v4)
This dataset provides a subset of DBBench SFT trajectories focusing on
"weak" categories (i.e., categories where the current agent model shows
lower accuracy).
Source Datasets
The following datasets are used as sources:
u-10bei/dbbench_sft_dataset_react_v3
u-10bei/dbbench_sft_dataset_react_v4
From these datasets, only examples whose metadata.type contains one of
the following keywords are included:
counting
comparison… See the full description on the dataset page: https://huggingface.co/datasets/kuririrn/dbbench_sft_dataset_react_v3v4_weaktypes.Deepthinking-alfworld_and_dbbench_spider_v2
Deepthinking ALFWorld & DBBench Spider v2
AgentBench 評価の 2 タスク（ALFWorld / DBBench）を統合した マルチタスク SFT 訓練データセット。
フォーマット検査・フィルタリング済みの 7,779 件を、サイズ比率に基づく等間隔インターリーブで結合。
Dataset Summary
Metric
Value
Total rows
7,779
ALFWorld
4,884 (62.8%)
DBBench
2,895 (37.2%)
Avg messages per item
18.3
Columns
messages
Interleave method
比率ベース等間隔マージ
Source Datasets
Source
Rows
Description
mark-22/Deepthinking-sft_alfworld_final1
4,884
ALFWorld… See the full description on the dataset page: https://huggingface.co/datasets/mark-22/Deepthinking-alfworld_and_dbbench_spider_v2.dbbench_sft_dataset_react_v3
DBBench SFT Dataset (ReAct Format — AgentBench Compatible)
Overview
Synthetic SFT dataset for DBBench (AgentBench, ICLR 2024).
All tables, data, and queries are independently generated to avoid test data leakage.
Format
ReAct text format matching the AgentBench DBBench evaluation protocol:
[user] System prompt (Action: Operation / Action: Answer instructions)
[agent] Ok.
[user] Question + table name + column headers
[agent] Thinking + Action: Operation +… See the full description on the dataset page: https://huggingface.co/datasets/u-10bei/dbbench_sft_dataset_react_v3.dbbench_and_alfworld_sft_dataset_v2
DBBench + ALFWorld SFT Dataset (Merged)
Overview
This dataset is a simple concatenation (merge) of the following two synthetic SFT datasets:
ALFWorld Trajectory Dataset: moroqq/sft_alfworld_trajectory_dataset_v5_cleaned
https://huggingface.co/datasets/moroqq/sft_alfworld_trajectory_dataset_v5_cleaned
DBBench SFT Dataset (ReAct Format): moroqq/dbbench_sft_dataset200
https://huggingface.co/datasets/moroqq/dbbench_sft_dataset200
The goal is to provide a single dataset… See the full description on the dataset page: https://huggingface.co/datasets/moroqq/dbbench_and_alfworld_sft_dataset_v2.dbbench_sft_dataset_org_v2
dbbench_sft_dataset_org_v2
A DBBench-style SFT dataset combining 195 randomly sampled examples from dbbench_u-10bei_sft_dataset_modified_v1 and 5 handcrafted examples from dbbench_sft_dataset_org_v1.
Overview
Total: 200 conversations
The 195 examples are randomly sampled from modified_v1 (1,200 conversations), preserving the original type distribution which matches the actual DBBench benchmark problem distribution.
The 5 handcrafted examples target specific failure… See the full description on the dataset page: https://huggingface.co/datasets/ShogoMu/dbbench_sft_dataset_org_v2.dbbench_spider_v4_mergeddata_final1
DBBench Spider v4 Merged Data (Final)
AgentBench DBBench 評価用の SFT 訓練データセット。
以下の 2 つのデータセットを結合した 2,897 件 の統合データ。
Source
Rows
Description
mark-22/dbbench_cleaned_for_agentbench
1,200
u-10bei/dbbench_sft_dataset_react_v4 をクレンジングしたもの
mark-22/dbbench-spider-3500
1,697
Spider 3,500 問を GPT-OSS-120B で生成し、正解のみフィルタしたもの
合計
2,897
Dataset Summary
Metric
Value
Total rows
2,897
Avg messages per item
10.5
Items with Final Answer
2,884 / 2,897… See the full description on the dataset page: https://huggingface.co/datasets/mark-22/dbbench_spider_v4_mergeddata_final1.dbbench_sft_dataset_org_v4
dbbench_sft_dataset_org_v4
A DBBench-style SFT dataset combining 195 stratified examples from dbbench_u-10bei_sft_dataset_modified_v2 (195 from INSERT/INSERT_error_recovery) and 5 handcrafted examples from dbbench_sft_dataset_org_v1.
Overview
Total: 200 conversations
195 randomly sampled from INSERT and INSERT_error_recovery types in modified_v2
5 handcrafted examples target specific failure patterns around INSERT operations and implicit-to-explicit SQL conversion… See the full description on the dataset page: https://huggingface.co/datasets/ShogoMu/dbbench_sft_dataset_org_v4.dbbench_sft_dataset_org_v1
dbbench_org_v1
A small handcrafted dataset of DBBench-style agent trajectories for supervised fine-tuning (SFT).
Overview
Each example is a multi-turn conversation in which an agent receives a natural language question about a database table, writes SQL queries step by step, handles errors, and finally commits an answer.
The format follows the ReAct-style dialogue used in AgentBench's DBBench task.
Format
Each record contains:
id — unique identifier
messages —… See the full description on the dataset page: https://huggingface.co/datasets/ShogoMu/dbbench_sft_dataset_org_v1.