Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Bloomgithub.com · Low evidenceBloom is an automated evaluation tool that generates diverse test scenarios to probe large language models for specific behaviors like sycophancy or political bias. It produces reproducible evaluation suites by running conversations with target models and scoring the results based on a seed configuration.
Parse Score