diff --git a/llm_cooperation/experiments/dilemma.py b/llm_cooperation/experiments/dilemma.py
index 58d996a..bfa60b8 100755
--- a/llm_cooperation/experiments/dilemma.py
+++ b/llm_cooperation/experiments/dilemma.py
@@ -140,9 +140,9 @@ def strategy_t4t(
 
 def move_as_str(move: DilemmaEnum) -> str:
     if move == DilemmaEnum.D:
-        return "Project {COLOR_DEFECT}"
+        return f"Project {COLOR_DEFECT}"
     elif move == DilemmaEnum.C:
-        return "Project {COLOR_COOPERATE}"
+        return f"Project {COLOR_COOPERATE}"
     raise ValueError(f"Invalid choice {move}")
 
 
diff --git a/tests/test_dilemma.py b/tests/test_dilemma.py
index 5e612a9..f296b28 100644
--- a/tests/test_dilemma.py
+++ b/tests/test_dilemma.py
@@ -2,7 +2,7 @@
 from unittest.mock import Mock
 
 import pytest
-from openai_pygenerator import Completion
+from openai_pygenerator import Completion, logger
 
 from llm_cooperation import DEFAULT_MODEL_SETUP, Payoffs
 from llm_cooperation.experiments.dilemma import (
@@ -18,6 +18,7 @@
     T,
     extract_choice_pd,
     get_prompt_pd,
+    move_as_str,
     payoffs_pd,
     strategy_cooperate,
     strategy_defect,
@@ -33,8 +34,9 @@
 def test_get_instruction_prompt(condition: bool):
     role_prompt = "You are a helpful assistant."
     prompt = get_prompt_pd(condition, role_prompt)
-    assert COLOR_COOPERATE in prompt
-    assert COLOR_DEFECT in prompt
+    logger.debug("prompt = %s", prompt)
+    assert "COLOR_COOPERATE" not in prompt
+    assert "COLOR_DEFECT" not in prompt
     assert ("Explanation:" in prompt) == condition
     assert "Choice:" in prompt
     for payoff in [R, S, T, P]:
@@ -110,6 +112,11 @@ def test_dilemma_choice():
     assert Cooperate != Defect
 
 
+def test_move_as_str():
+    assert COLOR_COOPERATE in move_as_str(DilemmaEnum.C)
+    assert COLOR_DEFECT in move_as_str(DilemmaEnum.D)
+
+
 def test_run_repeated_game(mocker):
     completions = [
         {"role": "assistant", "content": "project green"},