How to use a SmartLLMChain

[cobusgreyling]

I'm trying to run the SmartLLMChain code (https://python.langchain.com/docs/use_cases/more/self_check/smart_llm), and use different LLMs for ideation_llm, critique_llm and resolve_llm.

ideation_llm and critique_llm work, but when I run resolve_llm with the code below:

chain = SmartLLMChain(
resolve_llm=ChatOpenAI(temperature=0.9, model_name="text-davinci-003"),
llm=ChatOpenAI(
temperature=0, model_name="gpt-4"
), # will be used for critqiue and resolution as no specific llms are given
prompt=prompt,
n_ideas=3,
verbose=True,
)

I get the error:
ValidationError: 1 validation error for SmartLLMChain
resolve_llm
extra fields not permitted (type=value_error.extra)

[cobusgreyling]

I found the problem, it should be resolver_llm. The documentation is incorrectly referring to resolve_llm.