Harrison Chase
Harrison Chase
to handle cases where it does not generate an action/action input
https://arxiv.org/pdf/2207.05221.pdf should be a separate LLMchain with a prompt that people can add on at the end
A chain where you first call an LLM to do classification choose which subchain to call, then you call that specific subchain @sjwhitmore @johnmcdonnell were you guys going to work...
There is an open PR (https://github.com/hwchase17/langchain/pull/131) but its not complete. Would be nice to finish it up and add it in
something like ``` from langchain.chains.base import Chain class BaseWhileChain(Chain): def _call(self, inputs: Dict[str, str]) -> Dict[str, str]: state = self.get_initial_state(inputs) while True: outputs = self.inner_chain(state) if self.stopping_critera(outputs): return outputs else:...