Can I read Factors Causing LLM Hallucinations on EtoBox?

Factors Causing LLM Hallucinations by zhangyiang.scu is a document available to read on EtoBox.

What is Factors Causing LLM Hallucinations about?

This document discusses the concept of unlearning in large language models (LLMs), focusing on the removal of harmful responses, copyrighted content, and hallucinations. The authors argue that unlearning, which only requires negative examples, is a cost-effective and efficient alternative to traditional reinforcement learning from human feedback (RLHF) for aligning LLMs with human values. They present three scenarios where unlearning is beneficial and demonstrate that it can achieve better alignment perform

Author

zhangyiang.scu

Language

EN