Open
Description
I am working on some really cool technology, which is the splitting of the standard input into two file descriptors. We have the usual one, which is read by the python interpreter in order to execute code, and another which is read exclusively by the input() function.
The main use case of this is you can do things like put the input function in a for loop. Then the process waits for input from the llm on each iteration.
while True
# calculations etc
text=input("break?")
if text=="break":
break
This gives the LLM time to think between iterations.
Metadata
Metadata
Assignees
Labels
No labels