README.md

metadata

dataset_info:
  - config_name: spider
    features:
      - name: db_id
        dtype: string
      - name: question
        dtype: string
      - name: query
        dtype: string
      - name: query_toks
        list: string
    splits:
      - name: train
        num_examples: 7000
      - name: validation
        num_examples: 1034
  - config_name: reviewed_trajectories
    features:
      - name: db_id
        dtype: string
      - name: question
        dtype: string
      - name: gold_sql
        dtype: string
      - name: trajectory
        list:
          struct:
            - name: step
              dtype: int32
            - name: thought
              dtype: string
            - name: tool
              dtype: string
            - name: argument
              dtype: string
            - name: observation
              dtype: string
      - name: final_sql
        dtype: string
      - name: test_suite_passed
        dtype: bool
    splits:
      - name: train
        num_examples: 500
  - config_name: filtered_rollouts
    features:
      - name: db_id
        dtype: string
      - name: question
        dtype: string
      - name: rollout_sql
        dtype: string
      - name: execution_accuracy
        dtype: float32
    splits:
      - name: train
        num_examples: 1200
configs:
  - config_name: spider
    data_files:
      - split: train
        path: spider/train-*
      - split: validation
        path: spider/validation-*
  - config_name: reviewed_trajectories
    data_files:
      - split: train
        path: reviewed_trajectories/train-*
  - config_name: filtered_rollouts
    data_files:
      - split: train
        path: filtered_rollouts/train-*

AskTheDatabase Trajectories & Evaluation Benchmark 🗄️

Procedural interactive exploration trajectories and relational benchmark datasets for natural language to SQL translation.

This dataset repository powers the training and empirical evaluation of AskTheDatabase, enabling models to move beyond brittle single-turn SQL generation toward verifiable, multi-turn schema exploration.

---

Configurations & Subsets

1. reviewed_trajectories (Harness-Zero Distillation)

High-quality multi-turn trajectories curated and verified by K_SQL procedural rules:

• Tools executed: inspect_schema, sample_data, execute_sql.
• Target: Teaches models to actively inspect table schemas and categorical distinct values before committing to final queries, eliminating schema hallucinations.

from datasets import load_dataset

dataset = load_dataset("chrisjcc/ask-the-database", "reviewed_trajectories")
print(dataset["train"][0]["trajectory"])

2. filtered_rollouts (Condition 3 RLVR Warmup)

Rollout trajectories generated by student models and filtered strictly by execution accuracy and test suite rewards.

dataset = load_dataset("chrisjcc/ask-the-database", "filtered_rollouts")

3. spider (Normalized Relational Benchmark)

Normalized Spider text-to-SQL benchmark train (7,000 samples) and dev (1,034 samples) splits with aligned SQLite database schemas.

dataset = load_dataset("chrisjcc/ask-the-database", "spider")

---

Schema & Features

• db_id (string): Unique identifier of the target SQLite database.
• question (string): User natural language inquiry.
• gold_sql (string): Ground-truth reference SQL query.
• trajectory (list of structs): Step-by-step interactive actions, thoughts, and environment observations.
• test_suite_passed (bool): Whether the final SQL passes the dual-fixture mutation test suite.

---

Licensing & Attribution

• Spider benchmark: Yale LILY Lab (CC BY-SA 4.0).
• Curated trajectories & fixtures: Apache 2.0.