RSI-Jev
Shanghua-Gao/RSI-Jev
Typed-decision models (noul / choice / score) trained by a self-improving loop of AI agents — checkpoints, the code that produced them, and every version that failed.
- Stars
- 137
- Forks
- 12
Why it is listed
A research system whose AI-agent experiment loop builds Jev-style System One models returning probabilities over typed options rather than prose. Its current documented release also accepts images. Listed for the alternative-model research and released artifacts, without certifying self-improvement, latency, benchmark scores or full Jev API compatibility.