o1-mini results

llm_toolkit/logical_reasoning_utils.py

@@ -15,6 +15,7 @@ from sklearn.metrics import (
     f1_score,
     confusion_matrix,
 )
+from llm_toolkit.llm_utils import invoke_openai_api
 
 
 print(f"loading {__file__}")
@@ -520,11 +521,11 @@ def reasoning_with_openai(
 ):
     return invoke_openai_api(
         user_prompt.format(row["puzzle"], row["truth"], row["text"]),
+        system_prompt=system_prompt if using_system_prompt else None,
         max_tokens=max_tokens,
         model=model,
         base_url=base_url,
         temperature=temperature,
-        using_system_prompt=using_system_prompt,
     )