diff --git a/evals/registry/evals/05_scipaper_biochart.yaml b/evals/registry/evals/05_scipaper_biochart.yaml new file mode 100644 index 0000000000..c48ee82d52 --- /dev/null +++ b/evals/registry/evals/05_scipaper_biochart.yaml @@ -0,0 +1,9 @@ +biochart: + id: biochart.test.v1 + metrics: [accuracy] + description: Test the model's ability to understand the figure from literature. + +biochart.test.v1: + class: evals.elsuite.rag_match_fuzzy:RAGMatch + args: + samples_jsonl: 05_biochart/samples_single_gemini.jsonl \ No newline at end of file