llmsort v0.14.0 — Readings, not rankings: pairwise LLM ratio judgements fitted into cardinal scores with uncertainty, active pair selection, and explicit cost and provenance.

Sign in to join the discussion — your thebay.events account works here.