The paper examines whether adding syntactic and rhetorical structure to text can improve the prediction of incoherence in large language model outputs. Experiments show that plain text actually yields higher accuracy, as the added structural information conflicts with the models’ architectures. The authors also demonstrate that coherence assessment can help detect misleading content by applying zero‑shot experiments to a Brazilian disinformation dataset.
arXiv:2609.34240v2 Announce Type: replace-cross Abstract: Existing open-ended generation metrics measure likelihood, lexical diversity, or distributional similarity in generic representation space, y...
The study examines whether Dutch language models exhibit coherence‑illusion effects similar to human readers, using texts that refer back to earlier context with words like ‘again’ and ‘too’. Surprisal at the critical word aligns with human acceptability and eye‑tracking data, showing that models are more surprised by incoherent continuations unless a matching distractor is present. Attention entropy and an energy metric from associative‑memory literature reveal heads that behave differently under coherence versus incoherence, and ablating these heads demonstrates transfer effects across experiments, indicating a shared underlying mechanism.
The study evaluates eleven autoregressive transformer models on English agreement attraction scenarios using a surprisal-based approach. Results show that while transformers match human reading times for prepositional phrase configurations, they perform poorly on object‑extracted relative clauses, with predictions diverging across models and failing to capture human interference patterns. The authors argue that current transformers cannot adequately model human morphosyntactic processing and call for more rigorous, comprehensive testing to avoid misleading conclusions from limited syntactic setups.
The paper investigates how adjectival modifiers affect the semantic plausibility of events, using the Adept benchmark of 16,000 English sentence pairs that differ by a single adjective. Experiments show that sentence transformers, despite being conceptually suited to the task, underperform compared to models like RoBERTa. The authors provide an error analysis and discuss the implications of their findings for future work on balancing training and test data.
arXiv:2604. 25860v2 Announce Type: replace-cross Abstract: Machine-generated text (MGT) detection requires identifying structurally invariant signals across generation models, rather than relying on model-specific fingerprints.
The paper introduces a taxonomy for the term "AI slop," a label for low‑quality AI‑generated text, and proposes interpretable dimensions to assess it. Using expert interviews and span‑level annotation, the authors find that binary slop judgments are somewhat subjective yet correlate with latent factors like coherence and relevance. Their framework can be applied to both detection and preference tasks, offering insights into linguistic and stylistic contributors to quality, while noting that fully automated, scalable methods remain an open challenge.
arXiv:2608.30260v1 Announce Type: cross Abstract: While it is well-established that prosody carries crucial cues for syntactic structure, the degree and nature of correspondence between these two dom...
The paper introduces a reference-based framework to analyze coherence and diversity in open-ended text generation. It evaluates these properties by aligning them with human trajectories, comparing them to human continuations, and estimating their likelihood under a human reference distribution. Experiments show that diversity alignment and mean-based comparisons correlate with human quality ratings, while reference likelihood also associates positively, though results vary by configuration.
arXiv:2609.36214v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used by people whose first language is not English, yet these users have been shown to receive systematic...
arXiv:2606. 18922v1 Announce Type: cross Abstract: Figurative language and negation are two areas that challenge current language models, however, both are widely used throughout written and spoken language.
The paper presents a large‑scale empirical study of 284 interpretable linguistic features to detect AI‑generated English text across 30 language models and 10 text domains. It finds that classifiers using only linguistic features can reliably distinguish machine‑from human‑written text, but most indicators are highly context‑dependent. Only lexical richness measures consistently generalize across model families and domains, offering robust signals for detection.
arXiv:2610.28092v1 Announce Type: cross Abstract: Persuasive writing exploits cognitive biases, recruiting a reader's mental shortcuts to shape belief, often without the reader noticing. We call the...