ScottLogic / prompt-injection

Application which investigates defensive measures against prompt injection attacks on an LLM, with a focus on the exposure of external tools.
MIT License
11 stars 9 forks source link

Switch to using new langchain constructs #898

Open chriswilty opened 2 months ago

chriswilty commented 2 months ago

In #897 we upgraded langchain lib to latest 0.1.* version, but without modifying our code to use the latest, recommended functional constructs. This issue is for switching over to use the LangChain Expression Language (LCEL) instead of the older, and in some cases deprecated, class constructs.

Tasks

See the LCEL cookbook for further code examples.