I find in your paper there is one GPT-4 baseline to do direct agent planning. However, I can only find the GPT-4 testing scripts in Maze and Chess. GPT-4 scripts in other environments are missing. Meanwhile, when running gpt4_maze.py, it reports error after several iterations.
File "gpt4_maze.py", line 112, in act
return text_history+(Text(response_json['action'].strip() + "\n", True),)
AttributeError: 'list' object has no attribute 'strip'
I wonder whether you can help solve it. Thank you.
Hello,
I find in your paper there is one GPT-4 baseline to do direct agent planning. However, I can only find the GPT-4 testing scripts in Maze and Chess. GPT-4 scripts in other environments are missing. Meanwhile, when running gpt4_maze.py, it reports error after several iterations. File "gpt4_maze.py", line 112, in act return text_history+(Text(response_json['action'].strip() + "\n", True),) AttributeError: 'list' object has no attribute 'strip'
I wonder whether you can help solve it. Thank you.
Best, Yongchao