All products
Groundwork
Status: ResearchAn evaluation harness for grounded generation.
- Category
- Research
- Platforms
- CLI · Python · TypeScript
- Since
- 2026
- Domain
- labs.bahukhandi-labs.com/groundwork
Groundwork measures whether a model's claims are actually supported by the passages it retrieved. Open source, deterministic, and designed to be run in CI rather than admired in a paper.
Claim-level attribution scoring
Decomposes an answer into atomic claims and scores each one against the retrieved context, rather than scoring the answer as an undifferentiated blob.
Deterministic by default
Fixed seeds, pinned judge versions and content-hashed datasets, so a regression in your score means a regression in your system.
Runs in CI
A single command, machine-readable output, and a non-zero exit code when attribution drops below your threshold.
Built with
- Python
- TypeScript
- Rust
Documentation
Elsewhere in the lab