I am using Codex API for automated code completion. I am invoking Codex API with the following parameters:
model="code-davinci-002",
prompt=input_prompt,
temperature=0,
max_tokens=500,
top_p=1,
n=5, # number_of_suggestions
frequency_penalty=0,
presence_penalty=0,
stop=STOP_STR,
)```
But I am always getting the same `response` for all `top_k` results returned by the Codex API.
What is getting wrong here?