From Parameters to Prompts: Understanding and Mitigating the Factuality Gap between Fine-Tuned LLMs

Open in new window