Hi! Is there a function to "flush" all stream left from the llm? I'm trying to make it so when the llm respond with some specific text, we reprocess the whole llm call, but if the text is found mid sentence the previous stream is concatenated at the beggining. Any ideas?