Implicit Causality-biases in humans and LLMs as a tool for benchmarking LLM discourse capabilitie
Dialogue & Discourse · 15.07.2026 · 17 (2) · 1-52
DOI: 10.5210/dad.2026.201 ↗Kurz gefasst
Fachartikel in der Zeitschrift Dialogue & Discourse, online erschienen am 15.07.2026. Verfasst von Florian Kankowski und 3 weiteren Autor:innen.
Abstract (Originalsprache)
In this paper, we compare data generated with mono- and multilingual LLMs spanning a range of model sizes with data provided by human participants in an experimental setting to investigate well-established discourse biases. Beyond the comparison as such, we aim to develop a benchmark to assess the capabilities of LLMs with discourse biases as a robust proxy for more general discourse understanding capabilities. More specifically, we investigated Implicit Causality verbs, for which psycholinguistic research has found participants to display biases with regard to three phenomena:\ the establishment of (i) coreference relations (Experiment 1), (ii) coherence relations (Experiment 2), and (iii) the use of particular referring expressions (Experiments 3 and 4). With regard to coreference biases we found only the largest monolingual LLM (German Bloom 6.4B) to display more human-like biases. For coherence relation, no LLM displayed the explanation bias usually found for humans. For referring expressions, all LLMs displayed a preference for referring to subject arguments with simpler forms than to objects. However, no bias effect on referring expression was found, as opposed to recent studies investigating human biases.
Zitieren
APA
Kankowski, F., Solstad, T., Zarrieß, S., & Bott, O. (2026). Implicit Causality-biases in humans and LLMs as a tool for benchmarking LLM discourse capabilitie. Dialogue & Discourse, 17(2), 1-52. https://doi.org/10.5210/dad.2026.201
Vancouver
Kankowski F, Solstad T, Zarrieß S, Bott O. Implicit Causality-biases in humans and LLMs as a tool for benchmarking LLM discourse capabilitie. Dialogue & Discourse. 2026;17(2):1-52. doi:10.5210/dad.2026.201
BibTeX
@article{Kankowski2026,
author = {Kankowski, Florian and Solstad, Torgrim and Zarrieß, Sina and Bott, Oliver},
title = {Implicit Causality-biases in humans and LLMs as a tool for benchmarking LLM discourse capabilitie},
journal = {Dialogue & Discourse},
year = {2026},
volume = {17},
number = {2},
pages = {1--52},
doi = {10.5210/dad.2026.201}
}
EUVIMED