All projects

jev-as-llm

damienen/jev-as-llm

TypeSafe's Jev, a judgment model that was never trained to write, made to chat one word at a time. Browser-only, bring your own OpenRouter key.

TypeScript
Stars
0
Forks
0

Why it is listed

Experimental chat assembles text word by word through repeated Jev multiple-choice calls, showing alternative probabilities and replaying recorded runs without a key.

Related projects

  • jev-fanout-bench

    Measured: asking TypeSafe Jev N questions in one call bills the state once. 2,976 real requests, raw data, exact billing check.

    Shared topics: openrouter

  • here-we-go-jev

    Local playground and test bench for TypeSafe's Jev System One model: typed questions, calibrated-probability answers, and side-by-side comparison with an LLM baseline.

    Shared topics: openrouter

  • jev-router-playground

    Interactive playground for testing Jev model-routing decisions against OpenRouter models

    Shared topics: openrouter