Hey folks, for those on the technical side struggling with using LLMs - we've just open-sourced the solution we have been using internally to detect hallucinations (we work with highly regulated industries and can't afford having mistakes in data).
Just heard at the Paperless Lab Academy that AWS is developing a similar neurosymbolic architecture to launch it for their top-tier customers in the upcoming months - you can fork it from our GitHub repository and start using right now ;)
Enjoy and please send me your feedback.