diff --git a/evals/registry/data/3d_globe_movement/samples.jsonl b/evals/registry/data/3d_globe_movement/samples.jsonl new file mode 100644 index 0000000000..abac30e955 --- /dev/null +++ b/evals/registry/data/3d_globe_movement/samples.jsonl @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:48ce9896d1d4d17303d3a6c6da98d33448ae613d1fcf4be2ca61f4b07e84c584 +size 19162 diff --git a/evals/registry/evals/3d_globe_movement.yaml b/evals/registry/evals/3d_globe_movement.yaml new file mode 100644 index 0000000000..f52e1f8ec4 --- /dev/null +++ b/evals/registry/evals/3d_globe_movement.yaml @@ -0,0 +1,9 @@ +3d_globe_movement: + id: 3d_globe_movement.dev.v0 + description: Test the model's ability to keep track of position and orientation in a 3D environment, using the globe of planet earth as the test environment. + metrics: [accuracy] + +3d_globe_movement.dev.v0: + class: evals.elsuite.basic.includes:Includes + args: + samples_jsonl: 3d_globe_movement/samples.jsonl