adding expected threshold for Tau Retail

Author: xzrderek

eval_protocol/benchmarks/test_tau_bench_retail.py

@@ -108,6 +108,7 @@ def tau_bench_retail_to_evaluation_row(data: List[Dict[str, Any]]) -> List[Evalu
     ],
     rollout_processor=MCPGymRolloutProcessor(),
     rollout_processor_kwargs={"domain": "retail"},
+    passed_threshold={"success": 0.65, "standard_error": 0.02},
     num_runs=8,
     mode="pointwise",
     max_concurrent_rollouts=50,