Statistical Knowledge Assessment for Generative Language Models
share
Generative Language Models (GLMs) have demonstrated capabilities to store factual knowledge and answer queries efficiently. Given varying prompts, does a GLM consistently generate factually correct answers? In this paper, we introduce a statistical knowledge assessment framework guided by latent variables and the KaRR metric, which quantifies a model's knowledge by computing its continuous probability across diverse text forms. We conduct a comprehensive comparison of knowledge across 14 GLMs using our framework, including LLaMA, Alpaca, OPT, and others. Our statistical knowledge assessment encompasses 600 relation types and exhibits a strong correlation (0.43 Kendall's τ) with human evaluation. Our findings reveal that the knowledge in GLMs with the same backbone architecture adheres to the scaling law, and that tuning on instruction-following data may compromise the model's ability to generate factually correct text consistently.
READ FULL TEXT
