ml-evaluation
falsify
Fix the pass bar before the test runs. PRML is an open spec and CLI that hashes and timestamps evaluation criteria (SHA-256, RFC 3161) so anyone can check them later. Python, JS, Go, Rust.
- 8
- Python
- MIT
- 2 d ago
llm-eval-tracker
Live index of LLM evaluation tools and benchmarks, refreshed every 15 minutes from GitHub
- 7
- Python
- 3 wk ago
falsify-cookbook
Field manual for the PRML v0.1 specification. 13 patterns, 4 anti-patterns, 4 working examples. CC0.
- 1
- Python
- CC0-1.0
- 3 wk ago
Data-Science-Challenge-Coursera-Project-Loan-Default-Prediction
Data Science Challenge from Coursera Project : Loan Default Prediction
- 1
- Jupyter Notebook
- 2 y ago
prml-verify-action
Verify PRML (Pre-Registered ML Manifest) commitments in CI. Block merges on tampered or regressed eval claims.
- 0
- Python
- MIT
- 4 d ago
Redwood
Alignment-faking and adversarial-reasoning research artifacts inspired by Redwood Research. Behavioral probes and controlled failure-mode analysis.
- 0
- Python
- MIT
- 3 d ago
falsify-giskard
PRML pre-registration for Giskard scenario results: commit an eval claim to a SHA-256 before the run, then verify the result against it.
- 0
- Python
- MIT
- 4 d ago
falsify-inspect
PRML adapter for Inspect AI: record an eval's pass criteria in a hashed manifest and check eval logs against them. Listed in the Inspect extensions catalog. MIT.
- 0
- Python
- MIT
- 4 d ago
falsify-js
JavaScript reference implementation of the PRML (Pre-Registered ML Manifest) v0.1 specification. Byte-equivalent to the Python reference across 21 conformance vectors. Zero runtime dependencies.
- 0
- JavaScript
- MIT
- 4 d ago
falsify-integrity-index
Public scorecard of how 27 ML eval claims meet 9 PRML falsifiability criteria. CC0 data; MIT tooling.
- 0
- CC0-1.0
- 3 wk ago
ranklab
Controlled study of offline recommender model-selection stability across logging regimes and behavioral targets.
- 0
- Python
- Apache-2.0
- 3 mo ago


