Developing a Multi-Function Customer Service AI
A startup aims to build a single AI assistant for their customer service team. They need this assistant to perform three distinct functions: 1) Answer frequently asked questions based on a knowledge base, 2) Summarize long customer complaint emails into a few bullet points, and 3) Translate support requests from Spanish to English. Based on your understanding of how language models acquire new skills, describe the most effective strategy for preparing the fine-tuning data to create this single, multi-talented assistant.
0
1
Tags
Ch.4 Alignment - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Application in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Evaluating Multi-Task Fine-Tuning Strategies for AI Assistants
Developing a Multi-Function Customer Service AI
A development team is building a single language model intended to serve as a versatile corporate assistant. The model must be able to summarize internal reports, answer questions based on a company knowledge base, and draft professional emails. After an initial training phase, the team observes that the model is excellent at drafting emails but performs poorly on summarization and question-answering. Which of the following adjustments to their training process is most likely to create a single model that is proficient in all three tasks?