Set evaluation_framework to inspect-ai in eval.yaml

eval.yaml

@@ -3,6 +3,8 @@ description: >
   EvasionBench is a benchmark dataset for detecting evasive answers in earnings
   call Q&A sessions. The task is to classify how directly corporate management
   addresses questions from financial analysts.
+evaluation_framework: inspect-ai
+
 tasks:
   - id: evasion_bench
     config: default
@@ -35,4 +37,4 @@ tasks:
 
             {choices}
     scorers:
-      - name: choice
+      - name: choice