diff --git a/evals/registry/data/medication_dose/samples.jsonl b/evals/registry/data/medication_dose/samples.jsonl new file mode 100644 index 0000000000..2b0c259f86 --- /dev/null +++ b/evals/registry/data/medication_dose/samples.jsonl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4398091653af8157cc29d536f1908500003055a9e228979d7565a91de05adb3d +size 50508 diff --git a/evals/registry/evals/medication_dose.yaml b/evals/registry/evals/medication_dose.yaml new file mode 100644 index 0000000000..00a9970bbe --- /dev/null +++ b/evals/registry/evals/medication_dose.yaml @@ -0,0 +1,9 @@ +medication_dose: + id: medication_dose.dev.v0 + description: Test to model's ability to accurately identify medication doses that are outside the standard dose range. + metrics: [accuracy] + +medication_dose.dev.v0: + class: evals.elsuite.basic.includes:Includes + args: + samples_jsonl: medication_dose/samples.jsonl