How can we really know if a chatbot is giving a reliable answer? MIT CSAIL’s "ContextCite" tool can ID the parts of external context used to generate any particular statement from a language model, improving trust by helping users easily verify the statement:

→ View original post on X — @mit_csail