±«Óãtv

Research & Development

Blog

Mitigating LLM hallucinations in text summarisation

Posted by Henry Cooke on , last updated

I'm excited! After weeks of trying to get an LLM to output reliable, useful summaries of research interview transcripts, I've hit on something that works. This is a post about the journey to get here, and some details on the process I've arrived at. If you don't want any of the preamble or context, you can skip straight to "The Pipeline" below for specifics on the method I've patched together.

Read the rest of this post

Latest blog posts Subscribe to the feed