Large language models have significantly enriched the toolkit for computational humor research, particularly in the automated generation of jokes and puns. A key innovation, contextual embedding vectors, offers new opportunities to revisit and refine earlier hypotheses. Notably, Petrovic and Matthews (2013) proposed a joke generation model based on the scheme"I like my X like I like my Y, Z"(e.g."I like my ice like I like my dreams, crushed"). They suggested that joke hilarity increases with: a) frequent association of Z with X and Y, b) rarity of Z, c) ambiguity of Z, and d) meaning distance between X and Y. Building on this, Winters et al. (2019) proposed a set of metrics, based on Google Ngrams and Word2Vector. In this work, three out of their five metrics are revisited with word embeddings: obviousness, compatibility, and comparison. Another measure, symmetry, defined as closeness of Z to both X and Y, is introduced here for the first time. Two models were used to collect the embedding vectors (OpenAI text-embedding-3-small and MiniLM all-MiniLM-L6-v2) on three datasets: JokeJudger, Expunations, and rJokes. The last two datasets, Expunations, and rJokes, were expanded by adding paired sentences that captured the ambiguous expression at the core of each joke in its two different meanings. Results revealed that models trained on the proposed metrics performed poorly in predicting humor ratings: on JokeJudger, the best model achieved 57.1% accuracy, below the 61.5% baseline, while performance on Expunations and rJokes was even lower. Nevertheless, the symmetry metric seems consistently associated with higher-rated jokes, suggesting it may capture a necessary -though not sufficient- property of humor.
—In December 2023, Oxford English Dictionary selected rizz as the Word of the Year . This study examines the relationship between conceptual meanings of rizz and its source word charisma through constructing semantic networks. Based on authentic corpus data from COCA, NOW, and English-language social media platforms, semantic networks are constructed with ROST-CM6. The findings indicate that while charisma is broadly used across different contexts to describe intrinsic personal magnetism, often associated with leadership, rizz is more narrowly referred to attractiveness from romantic partners. Moreover, this study reveals that the rapid diffusion of rizz is attributed to its viral circulation on social media, broadcast by mainstream media, celebrity endorsement, and simplicity. In conclusion, this study demonstrates how semantic network analysis can elucidate the evolving meanings of neologisms and provide insights into the mechanisms underlying their dissemination.
Xiaozhou Hou· International Journal of Lan...· 0 citations
Abstract Large language models (LLMs) are increasingly used not only to retrieve information but also to generate language on behalf of users, raising important questions about how they mediate cultural identity and authorial voice. This study investigates that issue through a comparative analysis of five widely used systems – ChatGPT, Claude, Copilot, Gemini, and Perplexity – using the Inglehart–Welzel cultural map as an empirical framework. The analysis is conducted in two stages: first, a quantitative assessment of cultural value alignment based on model-generated responses to standardized survey items under neutral and country-specific prompting conditions; and second, a qualitative examination of linguistic features in selected excerpts of responses to explore how voice and stance are constructed. Representational alignment is measured as the distance between model outputs and national cultural profiles. The results show that baseline responses cluster around English-speaking and European cultural regions, indicating shared default priors, while country-specific prompting improves alignment, though performance is uneven across models and persistent gaps remain in several non-Western contexts. These patterns are conceptualized as cultural shadows, reflecting the influence of dominant cultural norms in model outputs. Drawing on Hyland’s (2005) account of stance and McKinley’s (2026) view of generative artificial intelligence (GenAI) as a rhetorical force, the findings suggest that these patterns extend beyond value alignment to the mediation of voice and self-representation, with differences across models indicating that AI-generated language may shape how culturally situated perspectives are expressed.
Ashkan Goudarzi, Aylar Naderi Zonouz· Digital Studies in Language...· 0 citations
The Zettelkasten ("card case"), introduced by the German social scientist Niklas Luhmann in 1981, is an external memory system composed of notes linked through explicit cross-references. A mesh of cross-references allows larger conceptual structures to emerge as the collection grows. Luhmann regarded his system as an indispensable combination of outboard memory and research assistant.
Luhmann was known for his exceptional productivity and attributed much of it to his Zettelkasten. Others have tried to emulate his system with varying degrees of success.
Luhmann's Zettelkasten was entirely paper-based; later times have seen efforts to computerize the design. Few seem to have achieved or even understood the levels of engagement that Luhmann wrote about: how his card collection supported fluid thinking, creativity, and serendipity.
This paper presents a framework for understanding how such a system can achieve Luhmann-level support, why and how so many users fall short of these ideals. It pulls together aspects of information theory, cognitive and psychological science, human interface, movie production practice, Topic Maps data modeling, and library science to illuminate key concepts underlying highly effective thinking support.
The work reported here is the outgrowth of four main threads dating back to the early 2000s. First was a series of browser bookmarks managers that were in essence lightweight zettelkastens that worked with bookmarks instead of notes [
Passin 2003
]. Second was a program to allow writing XML source text, including RDF, in a highly simplified and intuitive way that relies on a few simple conventions [
Passin 2007
]. Third was the development of a Topic Maps engine and various applications built on it. The fourth was extensive experience with systems modeling and enterprise architecture.
Together these threads lead to an understanding that modeling systems need to be highly adaptable to changes and reorganization, that more semantics is captured by the structure than by individual information items, and that a flexible and adaptive system needs to present the user with extreme simplicity and intuitive operation as far as possible.
This paper proposes that successful Zettelkasten implementations must provide very low cognitive friction so users do not feel impeded by their system. They should provide an outline-style view, and a user should be able to restructure the outline at any time in a convenient manner. Metadata is included by extremely simple markup that relies on a few conventions. Finally, though the underlying model is a Topic Map, most of the Topic Map machinery and details should be left implicit. This is essential to keep the cognitive friction low.
Even though the framework presented in this paper seems complex, implementations can be surprisingly simple. This is demonstrated by the author's third-generation Zettelkasten implementation.
T. Passin· Balisage Series on Markup Te...· 0 citations
Lexical semantic change (LSC) detection investigates changes in word meaning over time, focusing on language use at the lexical-semantic level from a diachronic perspective. The field has made significant progress over the past two decades, driven by the increased availability of multilingual benchmarks, notable performance improvements, and growing interdisciplinary applications. In this paper, we review the evolution of LSC models and benchmark constructions within the context of popular shared tasks. By categorizing LSC models into generations defined by key components, our investigation suggests that performance breakthroughs have been largely driven by advances in
semantic representations
, transitioning from count-based models to recent transformer-based approaches. Notably, transformer models have established themselves as state-of-the-art by integrating
Word-in-Context
tasks, which emphasize
semantic proximity in context
. Furthermore, we review substantial studies that primarily leverage diachronic word embeddings to explore political, social, and cultural contexts beyond the linguistic domain. Our work provides valuable insights for future model development and encourages further interdisciplinary exploration within digital humanities and social sciences.
Jing Chen, Emmanuele Chersoni, Dominik Schlechtweg et al.· Natural Language Processing· 0 citations
How do large language models (LLMs) organize moral knowledge? Models detect moral content broadly, but detection is a low bar. We ask whether they go further, distinguishing moral foundations from one another and organizing the relationships between them geometrically. We train six independent linear probes on open-weight language models, one per Moral Foundations Theory (MFT) category (care/harm, fair/cheat, lib/oppress, loy/betray, auth/subv, sanc/degrade), and examine how the resulting directions relate to each other in representation space. We find the directions neither collapse into a single moral detector nor isolate from one another. Rather, they span a near-maximal number of independent dimensions while sharing a positive common component. The shared component is the signature of integration, and it is moral-specific relative to a matched non-moral concept battery built identically (mean pairwise cosine 0.26 vs. 0.013). The geometry is consistent across architectures and scale and reaches its integration regime early in pre-training, well before probe accuracy saturates. The structure the model discovers shows no evidence of the individualizing/binding distinction predicted by Moral Foundations Theory (an underpowered test: only 20 candidate partitions exist) but rather reflects corpus statistics. Extending to moral dilemmas, each dilemma direction partially composes from its component foundations, at 2.7x a mismatched-pair baseline, while the majority of its variance encodes conflict-specific structure. The model represents moral tension itself, not a pre-resolved judgment.
When a large language model (LLM) writes Harry Potter fanfiction, it reliably produces fundamental elements of the Hogwarts universe, such as recognizable places and characters. Human-written Harry Potter fanfictions, however, typically include these fundamentals and much more, incorporating stylistically irregular content and relationship-diverse plotlines. This gap between LLM and human writing has been noted across a variety of domains. LLMs tend to produce"average"writing, while human writing contains more diverse content that covers a broader distribution. Existing work has shown the existence of this distributional"gap", but no work has proposed a systematic way to measure it. Our paper proposes a human-grounded framework that uses the empirical distribution of human writing on a topic to measure the distributional breadth of LLM-generated content on that same topic. We propose two metrics, LLM Coverage (LLM-Cov) and In-Boundary Rate (IBR), that separate the plausibility of LLM content from its distributional breadth. Across ideation and narrative tasks, we find that current LLMs produce plausible but narrow content that concentrates near the center of the human response space. Our framework can enable researchers to better assess the distributional breadth of LLM-authored content, which we term its"cultural reach".
What if pathology foundation models could do more with less? GigaPath-Flash and GigaTIME-Flash cut computational demands while maintaining strong performance, opening the door to larger studies and broader exploration. The post GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduAug 31, 2026
With millions of users across the world, Julia has been used to conduct cutting-edge research and to design new drugs, jet engines, heat pumps, and more.