Internet Texts Research Articles

Embers of autoregression show how large language models are shaped by the problem they are trained to solve

The widespread adoption of large language models (LLMs) makes it important to recognize their strengths and limitations. We argue that to develop a holistic understanding of these systems, we must consider the problem that they were trained to solve: next-word prediction over Internet text. By recognizing the pressures that this task exerts, we can make predictions about the strategies that LLMs will adopt, allowing us to reason about when they will succeed or fail. Using this approach-which we call the teleological approach-we identify three factors that we hypothesize will influence LLM accuracy: the probability of the task to be performed, the probability of the target output, and the probability of the provided input. To test our predictions, we evaluate five LLMs (GPT-3.5, GPT-4, Claude 3, Llama 3, and Gemini 1.0) on 11 tasks, and we find robust evidence that LLMs are influenced by probability in the hypothesized ways. Many of the experiments reveal surprising failure modes. For instance, GPT-4's accuracy at decoding a simple cipher is 51% when the output is a high-probability sentence but only 13% when it is low-probability, even though this task is a deterministic one for which probability should not matter. These results show that AI practitioners should be careful about using LLMs in low-probability situations. More broadly, we conclude that we should not evaluate LLMs as if they are humans but should instead treat them as a distinct type of system-one that has been shaped by its own particular set of pressures.

Новые формы прецедентного текста в современной интернет-коммуникации

The article examines the phenomenon of precedence in modern communication conditions, among the characteristics of which the authors note “text violence” as a process of perception and interpretation of information in the absence of the recipient’s conscious intention and need to perceive the content of the text. It is noted that within the framework of Internet communication, individual “text violence” is one of the sources of precedence. Thus, Internet communication is considered as a source of new precedent texts and as a tool for their forced introduction into the consciousness of communicants. The authors distinguish three levels of precedence: auto-precedent, social-precedent and national-precedent levels. As a precedent text, this article examines the erratives included in memes and, therefore, they have become widespread and entrenched in the speech of communicants. The authors classify erratives of this kind as a group of precedent erratives, while the second group consists of occasional erratives generated by the user directly during a communicative act. In order to identify the images generated in the minds of recipients when perceiving errative units used in precedent Internet texts, the authors conducted an associative experiment. A quantitative analysis of the experimental material demonstrates that most of the subjects are not only familiar with the precedent memes proposed to them, but also understand them in the form of erratives, offering their own interpretation. Only a small part of the participants made an attempt to edit the erratives according to the norms of the Russian language. Qualitative analysis of reactions obtained during free associative experiment allowed the authors to divide the resulting reactions into two groups: “acceptance” and “non-acceptance” by the communicants of errative precedent texts. In conclusion, it is stated that the errativization of modern communication and the acquisition of the status of precedent text by erratives through their popularization is an ongoing process.

THE POETIC NATURE OF INTERNET-TEXTS AS A NATIONAL CHARACTER TRAIT

The article analyzes texts produced through poetic tactics – one of the linguocultural units that reflects the national specifics of Kazakhstani Internet texts located in the virtual space. The Internet has also made it possible to develop poetic discourse, the authors of which can be any Internet user. Elements of poetic tactics used in the communications of Internet users model manifestations of national consciousness. Today, in the works of native speakers in the virtual space, it is common to find texts in the poetic form of the communicants’ own work or poetic texts that are not their own works, but partially or fully quoted from other sources, for example, from popular songs, poems of great folk poets, akyns. The goal is to consider the established concepts of poetic tactics in modern linguistics and study poetic texts that reflect the national characteristics of the Internet space. In the process of writing a work, methods of description, narration, and elements of linguocultural analysis techniques are used. They reveal the meaning of the poetic nature of communicants common in a certain folk culture, create the opportunity to determine their originality, and differentiate their functions. The choice of such tactics depends on the temperament, mentality, and values of a particular linguistic personality. As a result of linguoculturological and linguopersonological analysis of texts, it is concluded that poetic tactics in Internet texts depict ideas, the spirit of the people based on the national image of the world and form the text generation of the Kazakh virtual linguistic personality. Key words: poetic tactics, Internet text, communication, Internet commentary, picture of the world, virtual space.

Multi-word expressions for Russian L2 learners: corpora-based selection with expert verification

The article describes the experience of creating a corpus-based list of the most relevant multi-word expressions for Russian L2 learners, distributed across the levels of the Common European Framework of Reference for Languages (CEFR) from A1 to C1. Modern linguistic and cognitive research shows that our speech is patterned and largely consists of stable segments. This fact is supported by the linguodidactic idea of teaching not isolated language units but their combinations of different nature. However, the selection and ranking of multi-word expressions based on language proficiency levels is constrained by the difficulty of automatically extracting them from a corpus of texts and estimating their frequency, as well as disagreements in defining the boundaries, linguistic nature, and terminology of multi-word expressions. This article describes the experience of compiling a list of the most valuable fixed-type multi-word expressions from various sources: two types of existing CEFR-graded vocabulary lists for Russian L2 learners – lexical minimums for the TORFL (Test of Russian as a Foreign Language) system and Russian KELLY (KEywords for Language Learning for Young and adults alike); the most frequent n-grams from the RuFoLa – Russian L2 textbook corpus and from the Russian Web corpus of internet texts; list of discourse formulas from the «Pragmaticon» project. The CEFR level of each multi-word expression is predicted using the frequency-based Max Delta measure, and its effectiveness is subsequently validated through annotation by multiple experts. The resulting list of multi-word expressions contains 1645 entries from A1 to C1 levels. The proposed version of the list has been implemented into an automated text analysis system for learners of Russian as a Foreign Language and can be useful for a wide range of professionals in the preparation of educational content for foreign language learners. The suggested Max Delta measure has demonstrated a high degree of agreement with expert evaluations within proficiency levels A1-B1. This signifies the importance of further exploring its potential in addressing related practical tasks and in selecting language learning content derived for other languages.

Russian Logistics and Supply Chain Management: Challenges and Relevant Solutions

Relevance of the research. The article considers a forced paradigm shift and systemic restructuring of Russian logistics due to the introduction of sanctions barriers and partial trade blockade, which, in addition to restricting Russian exports and imports, threatens digital technological transformations in supply chain management and occurs against the backdrop of tightening international environmental requirements for business activities, including logistics.The purpose of the research is to review changes taking place in Russian logistics and to analyze corresponding solutions for supply chain management.Design, methods, information base. The methodological framework of the research consists of fundamentals of logistics and supply chain management, elements of economic systems digital development concept and basic principles of ecological economics. The main research methods are collecting information from secondary sources, practicing experts in-depth interviewing, specialists surveying, computer content and PESTEL analysis. The methodological basis is formed by the basics of logistics and supply chain management, elements of the concept of digital development of economic systems and basic principles of ecological economy. The information base of the research consisted of interview transcripts, survey results, bibliographic documents, and Internet text files.Findings. As a result of the work, external challenges and influencing factors are identified, internal processes, drivers and trends characteristic of Russian logistics and supply chain management today are identified and described. Relevant decisions made by Russian logisticians in the current period are systematized and evaluated. The flexibility and sustainability of supply chains was confirmed, as well as the high adaptability of Russian logistics to radical transformation of business environment and operating conditions.Originality/value of the research. The study showed that despite the deglobalization of Russian logistics, it remains possible to manage any cargo delivery to Russia from anywhere in the world. The material may be useful both to Russian logistics practitioners and to persons studying or teaching the discipline “Logistics”.

Образные единицы в коммуникативных практиках медиатизированного политического дискурса

The article is dedicated to the communicative features of mediatized political discourse, which remain insufficiently explored. Using the example of the use of figurative means of language in the genres of media-political communications, it is demonstrated that the mediatized form of communication possesses not only the basic functions of professional discourse but also additional functions: amplifying or softening content, saving linguistic means, emotional and attractive functions. Each of these functions plays a significant role in shaping audience perception and influencing public opinion on political issues. With the help of contextual and semantic analysis, as well as the corpus method, the following results were obtained: in the studied context, figurative units are important linguistic means used for the effective conveyance of specific political views and emotions. These units include figurative metaphors and comparisons, phraseology, proverb, precedent phenomena and colloquial expressions. The novelty of the research lies in the analysis of communicative features and functions of figurative units in various genres of mediatized political discourse: political talk shows, political interviews, and internet texts of politicians. The research results can be applied in the practice of media communication, communicative linguistics, and discourse theory.

The female nomination on the basis of marital status in Russian Internet texts: functional and linguo-ecological aspects

The article is devoted to the gendered lexical unit which constitutes initial abbreviation derived from the idiom «razvedenka s pricepom» (divorsed woman with a child) and is actively used by Russian speaking Internet users to denote a female person on the basis of her marital status. The article describes the history of the appearance of the unit, its derivational and semantical characteristics, semantic and contextual connections and its functioning specifics in the Internet texts. The research is conducted on the material of the author's corpus of the contexts of use including 1 255 speech products published in between 2014 and 2023 in the communicative space of such social platforms as (425 units), vk.com (336 units), YouTube (190 units), Yandex.Dzen (173 units). The author applies the methods of quantative, semantic, conceptual, derivational, contextual and discursive analyses. The article specifies that the gendered abbreviation is derived from the linguocynical idiom, having been used in the male speech long before the spread of the Internet. The research reveals that the abbreviation use underlines the artifact conceptual metaphor generated by the initial phraseological combination. The divorced mother deanimization is regularly strengthened by the abrriviation's contextual environment. It reveals that 1 175 contexts (93,62 %) contain verbalized speaker's conception of financial aspect in sexual, romantic or family relationship caused by the actualization of consumer strategy in interpersonal ineteraction and by the ubiquitous consumeristic philosophy. It proves that the use of the innovation RSP is a feature of male Internet users' speech (982 contexts; 78,24 %) and is not typical for female speech (270 contexts; 21,51 %). In female speech the abbreviation carries out primarily the metalinguistic function and much less frequently nominative and evaluative functions. In the Internet texts produced by male persons the nomination not only promotes expressing the divorced mother with a child negative evaluation but also facilitates deconstruction, deaxiologization and desacralization of the concept «Mother». This allows us to qualify the abbreviation as a linguo-perversion which is a special kind of language trauma which represents a speech phenomenon generating the traditional symbol ambivalence and leading towards worldview destruction.

Extracting intersectional stereotypes from embeddings: Developing and validating the Flexible Intersectional Stereotype Extraction procedure.

Social group-based identities intersect. The meaning of "woman" is modulated by adding social class as in "rich woman" or "poor woman." How does such intersectionality operate at-scale in everyday language? Which intersections dominate (are most frequent)? What qualities (positivity, competence, warmth) are ascribed to each intersection? In this study, we make it possible to address such questions by developing a stepwise procedure, Flexible Intersectional Stereotype Extraction (FISE), applied to word embeddings (GloVe; BERT) trained on billions of words of English Internet text, revealing insights into intersectional stereotypes. First, applying FISE to occupation stereotypes across intersections of gender, race, and class showed alignment with ground-truth data on occupation demographics, providing initial validation. Second, applying FISE to trait adjectives showed strong androcentrism (Men) and ethnocentrism (White) in dominating everyday English language (e.g. White + Men are associated with 59% of traits; Black + Women with 5%). Associated traits also revealed intersectional differences: advantaged intersectional groups, especially intersections involving Rich, had more common, positive, warm, competent, and dominant trait associates. Together, the empirical insights from FISE illustrate its utility for transparently and efficiently quantifying intersectional stereotypes in existing large text corpora, with potential to expand intersectionality research across unprecedented time and place. This project further sets up the infrastructure necessary to pursue new research on the emergent properties of intersectional identities.

Investor sentiment and stock returns: New evidence from Chinese carbon-neutral stock markets based on multi-source data

Combining internet text and stock trading data, we construct two investor sentiment proxies for 78 carbon-neutral stocks in Chinese stock markets based on the FinBERT model and principal component analysis. We then examine the effects of the two investor sentiment indices on carbon-neutral stock returns using dynamic panel models. We also investigate the mediating role of liquidity on the effects of investor sentiment. Our results reveal that both trading and internet sentiment negatively affect stock returns through the mediating effect of liquidity. This finding holds after robustness checks for dividing the sample into pre and post-epidemics. However, due to the effects of COVID-19, carbon-neutral stock returns are found to be more susceptible to investor sentiment.

The digital-mediated extensive reading on English Language learning of agriculture students

Reading is a skill that is initially taught in schools. Students are introduced to the alphabet and later learn to read texts. Reading Foreign languages is also introduced when students learn a foreign language. The student requires reading activities that include vocabulary learning. Extensive reading is one of the practical activities of reading. Nowadays, extensive reading activities include reading printed and digital text in e-books, journals, and internet and social media articles. In addition, digital extensive reading research on university students of agriculture is still rare. Thus, research on digital-mediated extensive reading in agricultural students’ needs to be conducted to investigate whether it improves their reading comprehension. This study involved 84 agricultural students majoring Agroecotechnology study program. The participants take an English course. The data for this study was taken from the pre-test and post-test scores and questionnaire. The data will be analyzed using descriptive and inferential statistics. The results of the study shows that the digital-mediated extensive reading effectively enhances the student's reading comprehension in English as academic purposes (EAP) setting, and the students express that the digital-mediated extensive reading provides flexibility, technology utilization, and reading.

ChatGPT offers an editorial on the opportunities for chatbots in dermatologic research and patient care.

ChatGPT is a chatbot developed by OpenAI, an artificial intelligence research laboratory, that is trained on massive-scale internet text data to understand a broad range of language styles and topics. As a mature, conversational chatbot, ChatGPT can respond to follow-up questions and produce coherent primary texts based on the user's request. We explore the opportunities and risks of integrating chatbots into dermatologic patient care and research while presenting ChatGPT's response to the same question.

ChatGPT-A Generative Pre-Trained Transformer

ChatGPT is an advanced conversational AI model developed by OpenAI. It is designed to engage in natural and coherent conversations with users, providing human-like responses to a wide range of topics and questions. It leverages deep learning technique. ChatGPT uses a combination of machine learning techniques, including deep learning and natural language processing, to understand and generate human-like text. The model has been trained on a vast amount of internet text data to ensure its ability to generate relevant and contextually accurate responses. ChatGPT has applications in customer service, virtual assistants, and other conversational interfaces, offering a powerful tool for natural language understanding and generation. It is most likely used to generate human like responses and makes the communication interactive.

Generating Contextual Variables From Web-Based Data for Health Research: Tutorial on Web Scraping, Text Mining, and Spatial Overlay Analysis.

Contextual variables that capture the characteristics of delimited geographic or jurisdictional areas are vital for health and social research. However, obtaining data sets with contextual-level data can be challenging in the absence of monitoring systems or public census data. We describe and implement an 8-step method that combines web scraping, text mining, and spatial overlay analysis (WeTMS) to transform extensive text data from government websites into analyzable data sets containing contextual data for jurisdictional areas. This tutorial describes the method and provides resources for its application by health and social researchers. We used this method to create data sets of health assets aimed at enhancing older adults' social connections (eg, activities and resources such as walking groups and senior clubs) across the 374 health jurisdictions in Catalonia from 2015 to 2022. These assets are registered on a web-based government platform by local stakeholders from various health and nonhealth organizations as part of a national public health program. Steps 1 to 3 involved defining the variables of interest, identifying data sources, and using Python to extract information from 50,000 websites linked to the platform. Steps 4 to 6 comprised preprocessing the scraped text, defining new variables to classify health assets based on social connection constructs, analyzing word frequencies in titles and descriptions of the assets, creating topic-specific dictionaries, implementing a rule-based classifier in R, and verifying the results. Steps 7 and 8 integrate the spatial overlay analysis to determine the geographic location of each asset. We conducted a descriptive analysis of the data sets to report the characteristics of the assets identified and the patterns of asset registrations across areas. We identified and extracted data from 17,305 websites describing health assets. The titles and descriptions of the activities and resources contained 12,560 and 7301 unique words, respectively. After applying our classifier and spatial analysis algorithm, we generated 2 data sets containing 9546 health assets (5022 activities and 4524 resources) with the potential to enhance social connections among older adults. Stakeholders from 318 health jurisdictions registered identified assets on the platform between July 2015 and December 2022. The agreement rate between the classification algorithm and verified data sets ranged from 62.02% to 99.47% across variables. Leisure and skill development activities were the most prevalent (1844/5022, 36.72%). Leisure and cultural associations, such as social clubs for older adults, were the most common resources (878/4524, 19.41%). Health asset registration varied across areas, ranging between 0 and 263 activities and 0 and 265 resources. The sequential use of WeTMS offers a robust method for generating data sets containing contextual-level variables from internet text data. This study can guide health and social researchers in efficiently generating ready-to-analyze data sets containing contextual variables.

How Generative AI Was Mentioned in Social Media and Academic Field? A Text Mining Based on Internet Text Data

How Generative AI Was Mentioned in Social Media and Academic Field? A Text Mining Based on Internet Text Data

Exploring the Role of Chinese Language and Literature in the Transmission of Traditional Culture by Combining the Method of Internet Text Analysis

Abstract This paper constructs a Bayesian network text recognition model based on the Bayesian network and explores the role of Chinese language literature in the dissemination of traditional culture by analyzing the embodiment of traditional culture in Chinese language literature network texts. The collection process of Chinese language and literature data in network text is analyzed from the perspective of textual data interaction. The information of node variables in a Bayesian network is used to determine the mutual relationship between Chinese language literature and traditional culture. The degree of interdependence between Chinese literature and traditional culture can be measured by combining mutual information. The results show that the correct rate of text recognition of the Bayesian text recognition model decreases slightly when the training samples are (100-300), but the correct rate always stays around 0.85, thus reflecting the effectiveness of the network recognition model in this paper. Chinese language literature has a certain role in the dissemination of traditional culture, which proves that Chinese language literature, as a carrier of traditional culture, can improve the dissemination speed of traditional culture. This study focuses on the integration of Chinese literature and traditional communication to improve a new vision.

Discursive formulas of response in Modern Greek

The article is dedicated to the study of discursive formulas of response in Modern Greek. Discursive formulas of response, i.e. idioms which are used in speech as a reaction to what was previously said by the interlocutor and which reproduce some of the phonetic properties of the previous utterance, constitute one of the most language-specific layers of phraseology and pose a difficulty for a correct decoding and translation. The aim of the article is to describe the little-studied material of Modern Greek formulas of response, to determine their types and specific properties and to identify typical communicative situations in which they are used. The study is based on the instances of use of such structures in literary works, internet texts, movies and series in Modern Greek, along with data from dictionaries. Methods of structural, semantic, contextual and comparative analysis are used. The findings of the study provide evidence that in Modern Greek there exist structures corresponding to the definition of discursive formulas of response. Eight types of such structures can be distinguished based on the type of the initiating utterance: reactions to a question; reactions to an address; reactions to an expression of agreement or disagreement; reactions to an expression of opinion or an evaluation; reactions to a factual statement; reactions to a greeting, a farewell or a well-wish; reactions to an expression of gratitude; reactions to a directive. It is shown, however, that the key role in every case is played by the actual lexical make-up of the initiating utterances. From the authors’ point of view, what is particularly interesting for the comparative analysis of discursive formulas of response in different languages is the fact that some Modern Greek formulas of response have the form of a curse or of an ill-wish.

Analyzing Sentiment with Self-Organizing Map and Long Short-Term Memory Algorithms

This research delves into the impact of Chat Generative Pre-trained Transformer, one of Open Artificial Intelligence Generative Pretrained Transformer models. This model underwent extensive training on a vast corpus of internet text to gain insights into the mechanics of human language and its role in forming phrases, sentences, and paragraphs. The urgency of this inquiry arises from Chat Generative Pre-trained Transformer emergence, which has stirred significant debate and captured widespread attention in both research and educational circles. Since its debut in November 2022, Chat Generative Pre-trained Transformer has demonstrated substantial potential across numerous domains. However, concerns voiced on Twitter have centered on potential negative consequences, such as increasedforgery and misinformation. Consequently, understanding public sentiment toward Chat Generative Pre-trained Transformer technology through sentiment analysis has become crucial. The research’s primary objective is to conduct Sentiment Analysis Classification of Chat Generative Pre-trained Transformer regarding public opinions on Twitter in Indonesia. This goal involves quantifying and categorizing public sentiment from Twitter’s vast data pool into three clusters: positive, negative, or neutral. In the data clustering stage, the Self-Organizing Map technique is used. After the text data has been weighted and clustered, the next step involves using the classification technique with LongShort-Term Memory to determine the public sentiment outcomes resulting from the presence of Chat Generative Pre-trained Transformer technology. Rigorous testing has demonstrated the robust performance of the model, with optimal parameters: relu activation function, som size of 5, num epoch som and num epoch lstm both at 128, yielding an impressive 95.07% accuracy rate.

Metaphor as Invective in the Genre of Internet Commentary: A Focus on German Political Discourse

This article presents the results of a linguo-stylistic analysis of a Germanlanguage commentary within the context of public political internet discourse. The analysis aims to identify distinctive features specific to this genre. Empirical material consists of over 200 comments on internet posts made by the German political party “Alternative for Germany,” selected through random sampling. The study applies both general scientific and linguistic methods, such as discourse analysis and contextual analysis. The relevance of this chosen topic is justified by the need for research in the field of internet linguistics, particularly in addressing genre-related issues. The goal is to determine, identify, and describe the characteristic features of commentary as an internet text. The article thoroughly examines the approaches taken by German linguists to define the concept of “invective,” encompassing lexical-semantic perspectives, speech act theory, and post-structural discourse analysis. Additionally, the analysis of the material reveals that invective is a hallmark of German political internet discourse, with manifestations in both explicit and implicit forms. The article discusses implicatures, arising through irony and metaphorization. The author concludes that metaphor can serve as a tool of persuasion. Metaphor, as a persuasive device, involves the intentional alteration of the mental representation of a given object based on the metaphorical image imposed by the manipulator. This typically includes exaggeration in some aspects of the metaphorically represented phenomenon while diminishing others.

Network text and problems of its linguo-expert research (results of the round table "Linguistic expert research on conflict in Internet communication")

This material provides an overview of the reports and speeches made at the round table meeting organized on June 29, 2023 as part of the VII International Scientific Conference "Language in the coordinates of the mass media" (St. Petersburg State University, St. Petersburg). Meetings of expert linguists, included in the program of the annual international conference, have been held since 2015 and involve not only philologists-linguists, but also practitioners interested in the problems of safe use of language in various situations of verbal communication to discuss topical issues of modern linguo-expert Russian practice. This year, the subject of discussion of the round table participants was speech conflicts in different genres of Internet communication and the problems of their study in modern linguoexpert practice. During the discussion, in the reports and speeches of the participants, possible methods and techniques for studying new conflict Internet texts of various genres were discussed in line with the linguistic pragmatic, stylistics, media communicative, linguo-expert and legal aspects.

Конфликт интерпретаций текста как следствие когнитивного диссонанса участников интернет-обсуждений

Latent cognitive dissonance in Internet disputers manifests itself as a conflict of text interpretations. This cognitive interpretation process of pre-textual cognitive dissonance often results in aggressive comments. The author analyzed interpretive Internet texts (comments) to reveal the cognitive patterns and prove the effect of pre-textual, i.e., cognitive, dissonance on the language and content of Internet comments. Addressee's external interpretive activity, represented in the text, is determined by internal cognitive processes, i.e., pre-textual presumptions and expectations. The study was based on the article entitled The people are multiplying "on earth", which was posted on a Runet information network resource on January 30, 2023. The article, which featured national demographic problems and possible solutions, yielded 189 comments that clearly illustrated the mechanism of conflict of interpretations. The methods of content and intent analyses revealed two cognitive patterns: the first one correlated with the intentional content of the trigger-text whereas the second pattern did not. Both patterns covered a number of cognitive models. The original text, subjectively interpreted by the addressee, triggered a wide range of comments that followed various cognitive patterns. As a result, the discussion unfolded as a conflict scenario, i.e., a struggle of attitudes and opinions, determined by the pre-textual cognitive dissonance of the disputers.