Assessing Language Models with Scaling Properties

Abstract: LLMs have primarily been evaluated with perplexity. While perplexity quantifies the most comprehensible prediction performance, it does not provide qualitative information on the success or failure of models. Another approach for evaluating LLMs is thus proposed, using the scaling properties of natural language. Five such tests are considered, with the first two accounting for the vocabulary population and the other three for the long memory of natural language. The following models were evaluated with these tests: n-grams, probabilistic context-free grammar (PCFG), Simon and Pitman-Yor (PY) processes, hierarchical PY, and neural LLMs. Only the neural LLMs exhibit the long memory properties of natural language, but to a limited degree. The effectiveness of every test of these models is also discussed.