How Prompt Caching Cut Our LLM Bills in HalfSlides from my Applied AI Conf talk on how Langdock used prompt caching to cut LLM costs, what breaks cache hits, and the tooling needed to keep savings visible in production.
Introduction of an XAI Evaluation Framework for Fact-CheckingAn open-source framework for human-centric XAI evaluation of human-AI collaborative disinformation detection. Published at ACM MAD '24.