Compression Aware Intelligence and the Origins of Hallucination and Identity

Contradish tests whether your LLM app gives consistent answers when the same question is phrased differently. It generates paraphrases of your inputs, runs them through your app, and detects contradictions in the outputs. You get back the conflict reason and fix and the CAI score

www.contradish.com

www.compressionawareintelligence.com

A CAI score is how stable a system’s reasoning stays when the meaning stays the same.