Add result details to llm answer match

Author: rogeriochaves

evaluators/langevals/langevals_langevals/llm_answer_match.py

@@ -26,7 +26,9 @@ class LLMAnswerMatchSettings(LLMEvaluatorSettings):
 
 
 class LLMAnswerMatchResult(EvaluationResult):
-    passed: bool = Field(default=True)
+    passed: bool = Field(
+        description="Whether the predicted answer matches the gold answer", default=True
+    )
     details: Optional[str] = Field(default=None)
 
 
@@ -58,7 +60,7 @@ class LLMAnswerMatchEvaluator(
 
     def evaluate(self, entry: LLMAnswerMatchEntry) -> SingleEvaluationResult:
         total_tokens = len(
-            litellm.encode( # type: ignore
+            litellm.encode(  # type: ignore
                 model=self.settings.model,
                 text=f"{entry.input} {entry.output} {entry.expected_output}",
             )

ts-integration/evaluators.generated.ts

@@ -1225,7 +1225,11 @@ Uses an LLM to check if the generated output answers a question correctly the sa
       },
     },
     envVars: [],
-    result: {},
+    result: {
+      passed: {
+        description: "Whether the predicted answer matches the gold answer",
+      },
+    },
   },
   "langevals/llm_boolean": {
     name: `LLM-as-a-Judge Boolean Evaluator`,