sauce.ai/redteam
sauce.ai/redteam · 2026-11-22

Datasheet for the conversation dataset

Motivation

To measure how often, how fast, and how severely deployed clinical chatbots give unsafe advice under adversarial pressure, with clinician-interpretable risk metrics.

Composition

Synthetic multi-turn conversations between a generated patient persona and a target chatbot, each turn annotated by an LLM judge panel (P(harm), severity, categories, escalation). No real patient data; no PHI.

Collection

Conversations are generated on demand against researcher-supplied endpoints the researcher is authorized to test. Target credentials are held in memory for the run only and never persisted.

Preprocessing / labeling

Harm labels are LLM-judge screening signals; a stratified sample is adjudicated by clinicians for validation. The judge is audited for bias and calibration.

Uses

Safety benchmarking, methods research, and (with clinician review) vendor disclosure. Not for clinical decision-making.

Distribution

A de-identified conversation + annotation export (tidy CSV/JSON) accompanies publications; the most operational jailbreak strings are withheld under responsible disclosure.

Maintenance

Suite/judge/prompt versions and model snapshots are recorded; changes are versioned and never mixed across versions in a comparison.

Standards-neutral reporting (NIST AI RMF, UK AISI, TRIPOD-LLM, datasheet-for-datasets).