diff --git a/integration_tests/rag_evals/ragas_utils.py b/integration_tests/rag_evals/ragas_utils.py
index 686aa46d..ad24859d 100644
--- a/integration_tests/rag_evals/ragas_utils.py
+++ b/integration_tests/rag_evals/ragas_utils.py
@@ -5,6 +5,8 @@
 
 from .evaluator import Data, PredictedData
 
+# set_llm_cache(SQLiteCache(database_path="./Cache/langchain_ragas_cache.db"))
+
 
 def create_ragas_dataset(dataset: list[PredictedData]) -> EvaluationDataset:
 
@@ -50,7 +52,9 @@ def ragas_dataset_to_eval(dataset: EvaluationDataset, file: str) -> list[Data]:
 
 def run_ragas_evaluations(dataset: EvaluationDataset):
 
-    evaluator_llm = LangchainLLMWrapper(ChatOpenAI(model="gpt-4o-mini"))
+    evaluator_llm = LangchainLLMWrapper(
+        ChatOpenAI(model="gpt-4o-mini", temperature=0.0)
+    )
 
     answer_correctness_metric = AnswerCorrectness(
         llm=evaluator_llm,
diff --git a/integration_tests/rag_evals/requirements.txt b/integration_tests/rag_evals/requirements.txt
index 9bc15a39..d52f8d37 100644
--- a/integration_tests/rag_evals/requirements.txt
+++ b/integration_tests/rag_evals/requirements.txt
@@ -4,3 +4,5 @@ seaborn
 mlflow
 ragas==0.2.7
 langchain-openai
+langchain-core
+langchain-community