Use passed LLM for default chain in MultiPromptChain (#4418)

Currently, MultiPromptChain instantiates a ChatOpenAI LLM instance for
the default chain to use if none of the prompts passed match. This seems
like an error as it means that you can't use your choice of LLM, or
configure how to instantiate the default LLM (e.g. passing in an API key
that isn't in the usual env variable).

langchain/chains/router/multi_prompt.py

@@ -9,7 +9,6 @@ from langchain.chains.llm import LLMChain
 from langchain.chains.router.base import MultiRouteChain
 from langchain.chains.router.llm_router import LLMRouterChain, RouterOutputParser
 from langchain.chains.router.multi_prompt_prompt import MULTI_PROMPT_ROUTER_TEMPLATE
-from langchain.chat_models import ChatOpenAI
 from langchain.prompts import PromptTemplate
 
 
@@ -54,9 +53,7 @@ class MultiPromptChain(MultiRouteChain):
             prompt = PromptTemplate(template=prompt_template, input_variables=["input"])
             chain = LLMChain(llm=llm, prompt=prompt)
             destination_chains[name] = chain
-        _default_chain = default_chain or ConversationChain(
-            llm=ChatOpenAI(), output_key="text"
-        )
+        _default_chain = default_chain or ConversationChain(llm=llm, output_key="text")
         return cls(
             router_chain=router_chain,
             destination_chains=destination_chains,