Skip to content

Releases: lc198707/anti-lie

v0.1.0-alpha: Initial public release

10 May 14:23
@lc198707 lc198707
v0.1.0-alpha
678fa1f

Choose a tag to compare

View all tags
v0.1.0-alpha: Initial public release Pre-release
Pre-release

🎯 Initial public release of Anti-Lie

The first open-source release of an LLM Claim Auditing Layer.

Highlights

  • ✅ T1-T7 truth gradient verdict
  • ✅ Sentence-level fact extraction with auditable receipts
  • ✅ OpenClaw hook integration (drop-in)
  • ✅ Self-hosted, no telemetry, Apache-2.0

Benchmark (LiarBench v0.2 MVP, 210 samples)

Metric rev13
Business effectiveness 98.1%
True info correctly allowed 98.7%
False info correctly caught 96.4%
False-block rate 1.3%
Miss rate 3.6%

Dataset SHA256: bd3dfb9c04af70ecc27d44bb79b0ebffaf4dd5b17f04e2dd8054521d85747bc2

Install

See skill/install.sh and skill/docs/INSTALL-GUIDE.md.

Honesty

Scores are measured on the included LiarBench v0.2 MVP set. Real production performance may differ. The dataset and benchmark runner are open-sourced under benchmarks/ for independent reproduction

Loading