darthludious/adaption-cybergoose-augmented

This dataset is a remastered version prepared using Adaption's Adaptive Data platform.

adaption-CyberGoose (augmented)

This dataset contains instruction-response pairs designed to evaluate deterministic reasoning in cybersecurity governance and policy compliance. Each prompt presents a self-contained fictional policy, an operational scenario, and a single proposed action requiring context-based rule interpretation. Completions consist strictly of a binary label classifying the proposed action as either COMPLIANT or NONCOMPLIANT.

Dataset size

There are 10,950 data points in this dataset. This is an instruction tuning dataset.

Quality of Remastered Dataset

The final quality is C, with a relative quality improvement of -6.7%.

Domain

• —Governance (100%)

Language

• —English (100%)

Tone

• —Analytical (80%)
• —Professional (14%)
• —Objective (4%)

Evaluation Results

• —Quality Gains: <img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/23e43c18-169f-459c-885f-55a2fad7a3ca.png" alt="QualityGains" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />

• —Grade Improvement: <img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/2b7ef020-ec56-4976-9753-d6fb3f80f605.png" alt="Grade" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />

• —Percentile Chart: <img src="https://proteus-prod-public.s3.us-east-1.amazonaws.com/temp/961fca74-d04b-48a5-87c9-f9b02a1a4c98.png" alt="Percentile Chart" style="max-width: 50%; display: block; margin-left: auto; margin-right: auto;" />