Most tested LLMs fail to reproduce human implicit causality biases in coreference, coherence, and referring-expression form, even when they show partial coreference effects.
Oliver Bott, Matthias Schrumpf, Jens Michaelis, and Torgrim Solstad
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Implicit Causality-biases in humans and LLMs as a tool for benchmarking LLM discourse capabilities
Most tested LLMs fail to reproduce human implicit causality biases in coreference, coherence, and referring-expression form, even when they show partial coreference effects.