mirror of
https://github.com/hwchase17/langchain
synced 2024-11-10 01:10:59 +00:00
25 lines
946 B
Python
25 lines
946 B
Python
"""**SmartGPT** chain is applying self-critique using the `SmartGPT` workflow.
|
|
|
|
See details at https://youtu.be/wVzuvf9D9BU
|
|
|
|
The workflow performs these 3 steps:
|
|
1. **Ideate**: Pass the user prompt to an `Ideation LLM` n_ideas times,
|
|
each result is an "idea"
|
|
2. **Critique**: Pass the ideas to a `Critique LLM` which looks for flaws in the ideas
|
|
& picks the best one
|
|
3. **Resolve**: Pass the critique to a `Resolver LLM` which improves upon the best idea
|
|
& outputs only the (improved version of) the best output
|
|
|
|
In total, the SmartGPT workflow will use n_ideas+2 LLM calls
|
|
|
|
Note that SmartLLMChain will only improve results (compared to a basic LLMChain),
|
|
when the underlying models have the capability for reflection, which smaller models
|
|
often don't.
|
|
|
|
Finally, a SmartLLMChain assumes that each underlying LLM outputs exactly 1 result.
|
|
"""
|
|
|
|
from langchain_experimental.smart_llm.base import SmartLLMChain
|
|
|
|
__all__ = ["SmartLLMChain"]
|