diff --git a/common/generative_configs/generative_config_long.json b/common/generative_configs/generative_config_long.json
new file mode 100644
index 0000000000..b85abdab8b
--- /dev/null
+++ b/common/generative_configs/generative_config_long.json
@@ -0,0 +1,8 @@
+{
+    "max_new_tokens": 256,
+    "min_new_tokens": 8,
+    "top_p": 0.9,
+    "temperature": 0.9,
+    "do_sample": true,
+    "num_return_sequences": 2
+}
\ No newline at end of file
diff --git a/response_selectors/llm_based_response_selector/generative_configs/default_generative_config.json b/response_selectors/llm_based_response_selector/generative_configs/default_generative_config.json
deleted file mode 100644
index 1edf84f186..0000000000
--- a/response_selectors/llm_based_response_selector/generative_configs/default_generative_config.json
+++ /dev/null
@@ -1,8 +0,0 @@
-{
-  "max_length": 60,
-  "min_length": 8,
-  "top_p": 0.9,
-  "temperature": 0.9,
-  "do_sample": true,
-  "num_return_sequences": 2
-}
\ No newline at end of file
diff --git a/response_selectors/llm_based_response_selector/generative_configs/openai-chatgpt.json b/response_selectors/llm_based_response_selector/generative_configs/openai-chatgpt.json
deleted file mode 100644
index 107e944b80..0000000000
--- a/response_selectors/llm_based_response_selector/generative_configs/openai-chatgpt.json
+++ /dev/null
@@ -1,7 +0,0 @@
-{
-  "max_tokens": 64,
-  "temperature": 0.4,
-  "top_p": 1.0,
-  "frequency_penalty": 0,
-  "presence_penalty": 0
-}
\ No newline at end of file
diff --git a/response_selectors/llm_based_response_selector/generative_configs/openai-text-davinci-003-long.json b/response_selectors/llm_based_response_selector/generative_configs/openai-text-davinci-003-long.json
deleted file mode 100644
index a109e4db88..0000000000
--- a/response_selectors/llm_based_response_selector/generative_configs/openai-text-davinci-003-long.json
+++ /dev/null
@@ -1,7 +0,0 @@
-{
-  "max_tokens": 128,
-  "temperature": 0.4,
-  "top_p": 1.0,
-  "frequency_penalty": 0,
-  "presence_penalty": 0
-}
\ No newline at end of file
diff --git a/response_selectors/llm_based_response_selector/generative_configs/openai-text-davinci-003.json b/response_selectors/llm_based_response_selector/generative_configs/openai-text-davinci-003.json
deleted file mode 100644
index 107e944b80..0000000000
--- a/response_selectors/llm_based_response_selector/generative_configs/openai-text-davinci-003.json
+++ /dev/null
@@ -1,7 +0,0 @@
-{
-  "max_tokens": 64,
-  "temperature": 0.4,
-  "top_p": 1.0,
-  "frequency_penalty": 0,
-  "presence_penalty": 0
-}
\ No newline at end of file
diff --git a/response_selectors/llm_based_response_selector/server.py b/response_selectors/llm_based_response_selector/server.py
index bc7fc64012..ab30ca81b5 100644
--- a/response_selectors/llm_based_response_selector/server.py
+++ b/response_selectors/llm_based_response_selector/server.py
@@ -58,7 +58,10 @@ def select_response_by_scores(hypotheses, scores):
 
 def select_response(dialog_context, hypotheses, human_uttr_attributes):
     try:
-        curr_prompt = PROMPT + "\nHypotheses:\n" + "\n".join([f'"{hyp["text"]}"' for hyp in hypotheses])
+        if "transformers" in GENERATIVE_SERVICE_URL:
+            curr_prompt = "Hypotheses:\n" + "\n".join([f'"{hyp["text"]}"' for hyp in hypotheses]) + "\n" + PROMPT
+        else:
+            curr_prompt = PROMPT + "\nHypotheses:\n" + "\n".join([f'"{hyp["text"]}"' for hyp in hypotheses])
         logger.info(f"llm_based_response_selector sends dialog context to llm:\n`{dialog_context}`")
         logger.info(f"llm_based_response_selector sends prompt to llm:\n`{curr_prompt}`")