Dawid-Skene consensus for multi-agent and multi-LLM disagreement: learn each agent's reliability and the true answer with calibrated confidence. Zero dependencies, node-free, TypeScript-first.
nodejs typescript mcp voting reliability consensus multi-agent calibration openai bayesian zero-dependency confusion-matrix agents em-algorithm dawid-skene label-aggregation anthropic llm-as-judge hallucination-mitigation crowd-aggregation
-
Updated
Jun 24, 2026 - TypeScript