http://
chat.langchain.dev An open source chatbot grounded to answer questions about LangChain documentation Blog Post: http://
blog.langchain.dev/langchain-chat There have been a lot of question/answering demos, but we believe we add some unique insights (covered in the blog)
@langchain
-
LangChain Chatbot: Open Source Documentation Q&A Tool
By
–
-
LangChain Wolfram Alpha Integration Showcased in New Video
By
–
An awesome video by @JavaFXpert and @dralandthompson highlighting the LangChain Wolfram Alpha integration that @nickscamara_ did last week!
-
Agent Response Parsing Improvement for Final Answers
By
–
Parsing of agent responses – h/t @krrish_dh Previously, the agent response expected the "final answer" to start on the same line as the "Final Answer" prefix. For longer answers, this sometimes didn't happen We now longer enforce that it be on the same line
-
Wikipedia Search Now Includes Page Attribution for Source Citation
By
–
Wikipedia source information – @rubensmau When you get a document from wikipedia search, it now comes with a `page` attribute (the page it was fetched from) This is needed when doing question/answering with sources (so you can cite the correct page)
-
SQL Chain Update: Row Limit Parameter Implementation
By
–
Updated SQL chain – @fpingham Similar to the above, the SQL chain would sometimes fail if the query generated too many rows. @fpingham added a parameter to help limit the number of rows that are returned https://
langchain.readthedocs.io/en/latest/modu
les/chains/examples/sqlite.html#choosing-how-to-limit-the-number-of-rows-returned
… -
LangChain introduces multistep SQL chain for database queries
By
–
New multistep SQL chain Sometimes there are too many SQL tables to put information about all of them into context. There's now a chain that: 1. selects which tables are relevant to the query
2. runs the previous SQL chain only over those tables https://
langchain.readthedocs.io/en/latest/modu
les/chains/examples/sqlite.html#sqldatabasesequentialchain
… -
Version 0.0.64 Release: Enhanced Robustness for Large Context
By
–
v0.0.64 a bunch of updates making it more robust to large context Updated API chain – @fpingham Updated SQL chain – @fpingham New multistep SQL chain
Parsing of agent responses – h/t @krrish_dh and Wikipedia source information – @rubensmau -
LangChain API Chain Update: Optimized JSON Response Generation
By
–
Updated API chain (and notebook) – @fpingham Sometimes, the API chain would fail if it generated too long of a json response. @fpingham changed the prompt to instruct to generate a minimal response He also added an example notebook! https://
langchain.readthedocs.io/en/latest/modu
les/chains/examples/api.html
… -
Dr-Doc-Search: Open Source PDF Conversation Pipeline
By
–
One project that people are sleeping on is `dr-doc-search` by @deskriders_twt A fully open source pipeline AND web app for conversing with PDFs
-
LangChain Adds Custom Prompt Support for LLMBash API QAEval
By
–
Custom prompts for LLMBash, API, and QAEval chains – @nickscamara_ Francisco Ingham Continuing the trend from yesterday, more chains now have the ability to pass in custom prompts! LLMBash: https://
langchain.readthedocs.io/en/latest/modu
les/chains/examples/llm_bash.html#customize-prompt
… QAEval: https://
langchain.readthedocs.io/en/latest/use_
cases/evaluation/question_answering.html#customize-prompt
…