Jawbreaker Scam Defense Data: Synthetic Training and Evaluation Sets

Synthetic and sanitized training and evaluation data for Jawbreaker, a local-first scam defense application. The dataset, created by build-small-hackathon, includes evaluation sets ranging from smoke checks to hard calibration suites. It was last updated on June 10, 2026.

Use Cases

• Training scam detection models based on synthetic suspicious text, email, or DM content.
• Evaluating model performance on scam defense tasks using provided evaluation sets.
• Benchmarking safety-focused NLP systems on calibration suites of varying difficulty.

Strengths

• Data is synthetic and sanitized, which may reduce privacy concerns.
• Includes structured evaluation sets for calibration and smoke checks.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

build-small-hackathon

Collection Method

Synthetic data generation.

Freshness

Last updated 2026-06-10 00:33:28