@mzsadiek@Saboo_Shubham_ Since none of the chunks semantically match with the query (“What is the longest sentence?”), then the LLM is prompted (by the developer) to reply no matches are found.
@mzsadiek@Saboo_Shubham_ What this application is doing is splitting your document into multiple small chunks, then performing a semantic similarity search (via embeddings) based on your query. Relevant chunks are injected into the LLM context.
@janseredynski What method are you using to resign? Doing so gives me integrity failures upon installation of IPA. I would like to think that this is possible since disabling SIP isn't ideal.