Attention Flows: Tracing LLM Conceptual Engagement via Story Summaries

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hicke, Rebecca M. M., Hamilton, Sil, Mimno, David, Kristensen-McLachlan, Ross Deans
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910110104682496
author Hicke, Rebecca M. M.
Hamilton, Sil
Mimno, David
Kristensen-McLachlan, Ross Deans
author_facet Hicke, Rebecca M. M.
Hamilton, Sil
Mimno, David
Kristensen-McLachlan, Ross Deans
contents Although LLM context lengths have grown, there is evidence that their ability to integrate information across long-form texts has not kept pace. We evaluate one such understanding task: generating summaries of novels. When human authors of summaries compress a story, they reveal what they consider narratively important. Therefore, by comparing human and LLM-authored summaries, we can assess whether models mirror human patterns of conceptual engagement with texts. To measure conceptual engagement, we align sentences from 150 human-written novel summaries with the specific chapters they reference. We demonstrate the difficulty of this alignment task, which indicates the complexity of summarization as a task. We then generate and align additional summaries by nine state-of-the-art LLMs for each of the 150 reference texts. Comparing the human and model-authored summaries, we find both stylistic differences between the texts and differences in how humans and LLMs distribute their focus throughout a narrative, with models emphasizing the ends of texts. Comparing human narrative engagement with model attention mechanisms suggests explanations for degraded narrative comprehension and targets for future development. We release our dataset to support future research.
format Preprint
id arxiv_https___arxiv_org_abs_2604_06416
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Attention Flows: Tracing LLM Conceptual Engagement via Story Summaries
Hicke, Rebecca M. M.
Hamilton, Sil
Mimno, David
Kristensen-McLachlan, Ross Deans
Computation and Language
Artificial Intelligence
Machine Learning
Although LLM context lengths have grown, there is evidence that their ability to integrate information across long-form texts has not kept pace. We evaluate one such understanding task: generating summaries of novels. When human authors of summaries compress a story, they reveal what they consider narratively important. Therefore, by comparing human and LLM-authored summaries, we can assess whether models mirror human patterns of conceptual engagement with texts. To measure conceptual engagement, we align sentences from 150 human-written novel summaries with the specific chapters they reference. We demonstrate the difficulty of this alignment task, which indicates the complexity of summarization as a task. We then generate and align additional summaries by nine state-of-the-art LLMs for each of the 150 reference texts. Comparing the human and model-authored summaries, we find both stylistic differences between the texts and differences in how humans and LLMs distribute their focus throughout a narrative, with models emphasizing the ends of texts. Comparing human narrative engagement with model attention mechanisms suggests explanations for degraded narrative comprehension and targets for future development. We release our dataset to support future research.
title Attention Flows: Tracing LLM Conceptual Engagement via Story Summaries
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2604.06416