English

Measuring Psychological Depth in Language Models

Computation and Language 2024-10-07 v2

Abstract

Evaluations of creative stories generated by large language models (LLMs) often focus on objective properties of the text, such as its style, coherence, and diversity. While these metrics are indispensable, they do not speak to a story's subjective, psychological impact from a reader's perspective. We introduce the Psychological Depth Scale (PDS), a novel framework rooted in literary theory that measures an LLM's ability to produce authentic and narratively complex stories that provoke emotion, empathy, and engagement. We empirically validate our framework by showing that humans can consistently evaluate stories based on PDS (0.72 Krippendorff's alpha). We also explore techniques for automating the PDS to easily scale future analyses. GPT-4o, combined with a novel Mixture-of-Personas (MoP) prompting strategy, achieves an average Spearman correlation of 0.51 with human judgment while Llama-3-70B with constrained decoding scores as high as 0.68 for empathy. Finally, we compared the depth of stories authored by both humans and LLMs. Surprisingly, GPT-4 stories either surpassed or were statistically indistinguishable from highly-rated human-written stories sourced from Reddit. By shifting the focus from text to reader, the Psychological Depth Scale is a validated, automated, and systematic means of measuring the capacity of LLMs to connect with humans through the stories they tell.

Keywords

Cite

@article{arxiv.2406.12680,
  title  = {Measuring Psychological Depth in Language Models},
  author = {Fabrice Harel-Canada and Hanyu Zhou and Sreya Muppalla and Zeynep Yildiz and Miryung Kim and Amit Sahai and Nanyun Peng},
  journal= {arXiv preprint arXiv:2406.12680},
  year   = {2024}
}

Comments

EMNLP 2024

R2 v1 2026-06-28T17:10:29.518Z