jevland

post commentary

Jev as a harness primitive: evals, verifiers, self-improvement

Notes from integrating Jev into a custom agent harness: beyond cheap classification, the interesting uses are continuous evals and monitoring, custom verifiers for long-horizon agents, and recursive self-improving harness loops; a guide is in progress.

Commentary; no figures yet.

commentary — Roundup, analysis or press. About this label

← Back to the directory