eval.yaml: declare evaluation_framework (none)

The Hub's benchmark spec asks eval.yaml for an evaluation_framework. This benchmark scores with its own metrics (accuracy, KL from gold, Brier, ECE), so it declares `none`, as InternScience/ResearchClawBench and VLABench/vlabench_primitive_ft_lerobot_video do. No task changes.

eval.yaml

@@ -4,6 +4,8 @@ description: >
   unstructured state and answers five typed questions about it at once, and
   every answer is a probability distribution, not a single label.
 
+evaluation_framework: none
+
 tasks:
   - id: accuracy
     config: all