add gpt4v metric_specs

src/helm/benchmark/run_specs/vlm_run_specs.py

@@ -124,6 +124,11 @@ def get_image2structure_metric_specs(
     ]
     return metric_specs + get_basic_reference_metric_specs()
 
+def get_gpt4v_originality_metric_specs() -> List[MetricSpec]:
+    return [
+        MetricSpec(class_name="helm.benchmark.metrics.gpt4v_originality_metrics.GPT4VOriginalityMetric", args={}),
+    ]
+
 
 ############################################################
 # VHELM run specs