3 ms·
Author of the original post here. We thought that most of the evaluations of LLMs have been around general-purpose and consumer-focused use cases. So, we decide
by palakiris 3y ago
Author of the original post here. We thought that most of the evaluations of LLMs have been around general-purpose and consumer-focused use cases. So, we decided to make our internal analysis public and publish our LLM evaluations on the boring enterprise use case of customer support automation.
If you are not already yawning, we evaluated the performance of popular LLMs on speed, cost, accuracy, hallucination, and other dimensions on proprietary datasets for customer support queries.