docs: added Constitutional AI references (#25553)

Added reference to the source paper.
This commit is contained in:
Leonid Ganeline 2024-08-20 05:00:58 -07:00 committed by GitHub
parent 4bd005adb6
commit d324fd1821
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -17,6 +17,8 @@
"source": [
"[ConstitutionalChain](https://api.python.langchain.com/en/latest/chains/langchain.chains.constitutional_ai.base.ConstitutionalChain.html) allowed for a LLM to critique and revise generations based on [principles](https://api.python.langchain.com/en/latest/chains/langchain.chains.constitutional_ai.models.ConstitutionalPrinciple.html), structured as combinations of critique and revision requests. For example, a principle might include a request to identify harmful content, and a request to rewrite the content.\n",
"\n",
"`Constitutional AI principles` are based on the [Constitutional AI: Harmlessness from AI Feedback](https://arxiv.org/pdf/2212.08073) paper.\n",
"\n",
"In `ConstitutionalChain`, this structure of critique requests and associated revisions was formatted into a LLM prompt and parsed out of string responses. This is more naturally achieved via [structured output](/docs/how_to/structured_output/) features of chat models. We can construct a simple chain in [LangGraph](https://langchain-ai.github.io/langgraph/) for this purpose. Some advantages of this approach include:\n",
"\n",
"- Leverage tool-calling capabilities of chat models that have been fine-tuned for this purpose;\n",
@ -324,7 +326,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.10.4"
"version": "3.10.12"
}
},
"nbformat": 4,