compitum
What it is
compitum is a geometrically-aware LLM router that uses SPD metric learning and Lyapunov-inspired control for constraint-first model selection. It provides a CLI interface for routing prompts and includes rigorous property-based and mutation testing for reproducibility.
The problem it solves
Standard LLM routers lack geometric awareness and rigorous stability guarantees, leading to inconsistent or suboptimal model selection.
What you could use it for
- Routing complex mathematical proofs to specialized LLMs
- Automated model selection based on geometric constraints
- Reproducible evaluation of LLM routing strategies
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.