diff --git a/docs/extensions/extensions.yml b/docs/extensions/extensions.yml index d2994dd51c..2ffea66beb 100644 --- a/docs/extensions/extensions.yml +++ b/docs/extensions/extensions.yml @@ -70,6 +70,12 @@ author: "[Debu Sinha](https://github.com/debu-sinha)" categories: ["Analysis"] +- name: "[Inspect falsify](https://github.com/studio-11-co/falsify-inspect)" + description: | + Pre-register an evaluation's success criteria (metric, threshold, dataset hash, seed) as a tamper-evident PRML manifest before the run, then verify eval logs against the locked claim via installed hooks. + author: "[Cüneyt Öztürk](https://falsify.dev)" + categories: ["Analysis"] + - name: "[CJE](https://github.com/cimo-labs/cje)" description: | Calibrated judge evaluation — calibrate model-graded scorer accuracy using causal inference with optional oracle labels.