v1.0.4
NVIDIA/garakv1.0.4Jul 17, 2026by virattt
AI Summary
A feature update for the garak LLM evaluation framework adding scoring capabilities, reproducibility controls, and improved reporting metrics.
Key Highlights
- Added rubric-based partial-credit scoring and contradiction detection
- Added reproducible sampling, quick runs, and configurable models
- Improved eval reporting with metadata, latency, and failure metrics
New Features
- Rubric-based partial-credit scoring
- Contradiction detection
- Reproducible sampling
- Quick runs
- Configurable models
- Configurable concurrency and timeouts
- Improved eval reporting (metadata, latency, failure metrics)
Full Release Notes
### Changes - Add rubric-based partial-credit scoring and contradiction detection - Add reproducible sampling, quick runs, configurable models, concurrency, and timeouts - Improve eval reporting with metadata, latency, and failure metrics ### Fixes - Fix inaccurate finance evaluation answers and rubric criteria