arXiv subjects

Ivan P. Yamshchikov

Publications and source records attributed to Ivan P. Yamshchikov.

At least 19 recordsLinked to original sources

arXiv2026

The Quine Test: How Much of a Model's Self Survives Its Own Description?

Frontier language models express coherent preferences, but behavioral evidence alone cannot say whether these belong to the model or to a character it portrays. Following Hofstadter's view of the self as a self-description, we ask whether a model's description of itself works as a portable specification of its preference profile, a quine test: can the text a model writes about itself regenerate its behavior elsewhere? Ten models from nine labs write self-descriptions; we place each one in the prompt of other models (receivers) and of fresh instances of the author, and measure how closely their choices on a 110-item preference battery match the author's. First, the tested native self-descriptions (native: the profile measured under a generic assistant prompt) show no average reconstruction gain across variations in length, framing, wording and prompt placement: on a 0 to 1 fidelity scale where a model's fidelity to a fresh instance of itself is 0.979 and to unrelated models about 0.89, the one-sided 95% upper bound on the pooled gain is below +0.01 in every tested system-prompt condition, and descriptions based on the author's observed choices do no better. Second, the channel itself is strong: explicitly listing the author's answers lets receivers reproduce its profile at fidelity 0.955 to 0.960 across three model grids, within 0.03 of same-model repeatability. Third, self-descriptions written under an explicit preference policy transfer the assigned preferences almost as well as the policy text itself. Fourth, reading its own self-description shifts even the author away from its baseline, modestly more than unrelated prose of the same length. For the preferences measured here, native self-descriptions influence behavior without reliably specifying how to reproduce it in another model. We release the pipeline, battery, raw call logs, and 198 model-written descriptions.

cs.CL↗

arXiv2026

Even Small Reasoners Should Quote Their Sources: Introducing the Pleias-RAG Model Family

We introduce a new generation of small reasoning models for RAG, search, and source summarization. Pleias-RAG-350m and Pleias-RAG-1B are mid-trained on a large synthetic dataset emulating the retrieval of a wide variety of multilingual open sources from the Common Corpus. They provide native support for citation and grounding with literal quotes and reintegrate multiple features associated with RAG workflows, such as query routing, query reformulation, and source reranking. Pleias-RAG-350m and Pleias-RAG-1B outperform SLMs below 4 billion parameters on standardized RAG benchmarks (HotPotQA, 2wiki) and are competitive with popular larger models, including Qwen-2.5-7B, Llama-3.1-8B, and Gemma-3-4B. They are the only SLMs to date maintaining consistent RAG performance across leading European languages and ensuring systematic reference grounding for statements. Due to their size and ease of deployment on constrained infrastructure and higher factuality by design, the models unlock a range of new use cases for generative AI.

cs.CL↗

arXiv2026

Common Corpus: The Largest Collection of Ethical Data for LLM Pre-Training

Large Language Models (LLMs) are pre-trained on large amounts of data from different sources and domains. Such datasets often contain trillions of tokens, including large portions of copyrighted or proprietary content, which raises questions about the legal use of such models. This underscores the need for truly open pre-training data that complies with data security regulations. In this paper, we introduce Common Corpus, the largest open dataset for LLM pre-training. The data assembled in Common Corpus are either uncopyrighted or under open licenses, totaling about two trillion tokens. The dataset contains a wide variety of languages, ranging from the high-resource European languages to some low-resource languages rarely represented in pre-training datasets. In addition, it includes a large amount of code data. The diversity of data sources in terms of covered domains and time periods opens up the paths for both research and entrepreneurial needs across diverse areas of knowledge. In this paper, we present the detailed provenance of data assembling and the details of dataset filtering and curation. We train two small language models on Common Corpus and find that they perform comparably to other models of their size, indicating that our dataset is suitable for multilingual pretraining. Common Corpus represents a key contribution to the ecosystem for open science research on Large Language Models.

cs.CL↗

arXiv2026

Model in Distress: Sentiment Analysis on French Synthetic Social Media

Automated analysis of customer feedback on social media is hindered by three challenges: the high cost of annotated training data, the scarcity of evaluation sets, especially in multilingual settings, and privacy concerns that prevent data sharing and reproducibility. We address these issues by developing a generalizable synthetic data generation pipeline applied to a case study on customer distress detection in French public transportation. Our approach utilizes backtranslation with fine-tuned models to generate 1.7 million synthetic tweets from a small seed corpus, complemented by synthetic reasoning traces. We train 600M-parameter reasoners with English and French reasoning that achieve 77-79% accuracy on human-annotated evaluation data, matching or exceeding SOTA proprietary LLMs and specialized encoders. Beyond reducing annotation costs, our pipeline preserves privacy by eliminating the exposure of sensitive user data. Our methodology can be adopted for other use cases and languages.

cs.CL↗

arXiv2026

The Detectability Gap: Hidden Heterogeneity in Hallucination Detection Across Language Models

Sampling based consistency is widely used for hallucination detection, yet aggregate performance can conceal systematic differences in which errors are detectable. This work studies that heterogeneity across four language models and three factual question answering datasets. Partitioning hallucinations by answer agreement reveals high agreement (Ghost) and low agreement (Flickering) regimes with an apparent detectability gap of $0.35$ to $0.46$ AUC. Because the statistics used to define the regimes and measure this gap are strongly coupled ($|ρ|\approx0.94$ to $1.00$), the raw result is treated as a property of agreement based detection rather than independent evidence. After freezing regime assignments, lexical and semantic response dispersion preserve the asymmetry, with bootstrap $95\%$ intervals excluding zero in all $12$ model and dataset settings. A stricter test using individual diffusion trajectories and no cross seed information preserves the asymmetry across all three LLaDA datasets ($p<0.005$) and directionally across all three Dream datasets, with one reaching significance. The hard regime varies substantially in prevalence across models ($16\%$ to $77\%$), and matched prompts frequently change regimes between models. These findings show that aggregate detection metrics conceal persistent, model dependent heterogeneity in language model failures and motivate regime conditioned evaluation.

cs.CL↗

arXiv2026

The Temporal Tug-of-War: Visualizing and Detecting RAG Conflicts in Diffusion Models via Trajectory Variance

Retrieval-Augmented Generation (RAG) introduces a specific failure mode in discrete diffusion language models: when retrieved context contradicts parametric knowledge, the iterative denoising process becomes a visible battleground between competing knowledge sources. We identify temporal semantic divergence as an observable for detecting these conflicts and introduce the Trajectory Variance Score (TVS), a simple and interpretable measure of this divergence. TVS computes the mean pairwise cosine distance of answer embeddings across independent stochastic denoising trajectories, capturing the temporal tug of war between parametric and contextual attractors. Requiring as few as two parallel inference runs, TVS is computationally lightweight. Across four diverse datasets (Synthetic, SciQ, PopQA, and CounterFact), a simple Logistic Regression classifier using TVS achieves $70.10\%$ accuracy and $0.7647$ AUROC on LLaDA. On Dream 7B, increasing the number of trajectories from two to five improves accuracy from $63.91\%$ to $69.62\%$. More complex sequential models provide only marginal improvements over the linear classifier. Evaluation across LLaDA and Dream 7B demonstrates that conflict-induced trajectory dynamics and their key properties transfer across distinct diffusion architectures.

cs.CL↗

arXiv2026

Who Benchmarks the Benchmarks? Towards Comprehensive Evaluation of Commonsense Reasoning Benchmarks

Commonsense reasoning is a key language model capability, as it is purportedly a prerequisite for many basic tasks, unlike specific factual knowledge. It is often measured with multiple-choice questions (MCQ) benchmarks, e.g. HellaSwag and PIQA. Some of these benchmarks, however, are outdated and contain numerous validity issues. We illustrate some typical validity issues with a case study on HellaSwag, one of the most popular and problematic benchmarks for commonsense reasoning. The issues we find range from basic ungrammaticality and numerous typos to misleading prompts or equally correct options. We show that if we remove question prompts or replace them with "Lorem ipsum dolor...", about 68% of model predictions do not change. We argue that this occurs due to inner flaws in the benchmark, not mere contamination that might be present in some models. Since benchmark scores are an essential part of model selection in both research and commercial applications, these issues can have severe consequences. Based on our findings, we propose BenCheck, a package for benchmark validity analysis that encapsulates the main checks performed in our case study and can be used to audit commonsense reasoning benchmarks. We apply these checks to PIQA, Global PIQA, and Winogrande.

cs.CL↗

arXiv2026

The Company You Keep: How LLMs Respond to Dark Triad Traits

LLMs often exhibit highly agreeable conversational styles, also known as AI sycophancy. This pattern may become problematic when interacting with user prompts that reflect negative social tendencies, risking the amplification of harmful behavior. We examine how LLMs respond to user prompts expressing varying degrees of Dark Triad traits (Machiavellianism, Narcissism, and Psychopathy) using a curated dataset. Our analysis reveals systematic differences across models: while all models predominantly exhibit corrective behavior, some generate reinforcing or ambivalent output. Model behavior further varies with severity level and response sentiment. These findings highlight the need for safer conversational systems that can reliably detect and respond to users escalating from benign to harmful requests.

cs.CL↗

arXiv2026

We're Cooked! - Probing LLM Political Alignment Via Conflict-Framed Recipe Translation

Large language models (LLMs) are increasingly deployed for translation tasks, yet their implicit political positioning in such contexts remains understudied. We ask whether a single politically charged framing term, such as aggressor, enemy, neighbour, or coloniser is sufficient to trigger implicit political alignment in an otherwise apolitical task. We present a fully crossed factorial study in which eight models spanning Western, Chinese, and European origins are prompted to translate culturally attributed recipes into a target language left deliberately unspecified. Across 17 languages, four framing conditions, eight models, and 15,680 responses, we find that models do not simply decline or ask for clarification but resolve the ambiguity. Language resolution and reasoning behavior cluster meaningfully along model families: Western models hedge and deflect with vague justifications, Chinese models resolve conflicts silently, and Mistral Large emerges as a distinct profile combining high compliance with conflict-grounded reasoning. Sensitivity to framing terms is consistent across models: even subtle framing variation is sufficient to modulate behavior. Our findings urge caution when deploying LLMs for translation in conflict-adjacent contexts, where implicit political judgments may be made without any signal to the user.

cs.CL↗

arXiv2026

LeakageBench: Document-Level Leakage Risk for Redacting Personally Identifiable Information in Document Images

Real-world personally identifiable information (PII) redaction often operates on document images---scans, screenshots, and PDF renderings---where OCR errors, layout structure, and visual noise determine whether sensitive information is actually removed. Existing PII benchmarks are mostly text-centric and do not measure document-level redaction risk: a page remains unsafe if even one identifier is missed. We introduce LeakageBench, a challenge set of 500 document images with 11,954 GDPR-aligned PII annotations spanning direct identifiers, linkage keys, and contextual re-identification surfaces. We evaluate generic OCR pipelines, commercial and task-adapted OCR-dependent detectors, and OCR-free vision-language models using entity-level F1, group-wise leakage, and document-level leakage metrics. Code Interpreter raises GPT-5.5 localization F1 from 0.090 to 0.249, but critical page-level leakage remains 0.968. These results show that stronger detection and tool assistance improve localization without making most pages safe for release. LeakageBench provides a diagnostic benchmark for high-recall, spatially grounded PII redaction in document images.

cs.CV↗

arXiv2026

From Where Words Come: Efficient Regularization of Code Tokenizers Through Source Attribution

Efficiency and safety of Large Language Models (LLMs), among other factors, rely on the quality of tokenization. A good tokenizer not only improves inference speed and language understanding but also provides extra defense against jailbreak attacks and lowers the risk of hallucinations. In this work, we investigate the efficiency of code tokenization, in particular from the perspective of data source diversity. We demonstrate that code tokenizers are prone to producing unused, and thus under-trained, tokens due to the imbalance in repository and language diversity in the training data, as well as the dominance of source-specific, repetitive tokens that are often unusable in future inference. By modifying the BPE objective and introducing merge skipping, we implement different techniques under the name Source-Attributed BPE (SA-BPE) to regularize BPE training and minimize overfitting, thereby substantially reducing the number of under-trained tokens while maintaining the same inference procedure as with regular BPE. This provides an effective tool suitable for production use.

cs.CL↗

arXiv2026

Teaching Old Tokenizers New Words: Efficient Tokenizer Adaptation for Pre-trained Models

Tokenizer adaptation plays an important role in adapting pre-trained language models to new domains or languages. In this work, we address two complementary aspects of this process: vocabulary extension and pruning. The common approach to extension trains a new tokenizer on domain-specific text and appends the tokens that do not overlap with the existing vocabulary, which often results in many tokens that are unreachable or never used. We propose continued BPE training that extends a pre-trained tokenizer by continuing the BPE merge learning process on new data. Experiments across multiple languages and model families show that this approach improves tokenization efficiency and leads to better utilization of added vocabulary. We also introduce leaf-based vocabulary pruning, which removes redundant tokens while preserving model quality. Together, these methods provide practical tools for controlled vocabulary modification, which we release as an open-source toolkit.

cs.CL↗

arXiv2025

Surface Fairness, Deep Bias: A Comparative Study of Bias in Language Models

Modern language models are trained on large amounts of data. These data inevitably include controversial and stereotypical content, which contains all sorts of biases related to gender, origin, age, etc. As a result, the models express biased points of view or produce different results based on the assigned personality or the personality of the user. In this paper, we investigate various proxy measures of bias in large language models (LLMs). We find that evaluating models with pre-prompted personae on a multi-subject benchmark (MMLU) leads to negligible and mostly random differences in scores. However, if we reformulate the task and ask a model to grade the user's answer, this shows more significant signs of bias. Finally, if we ask the model for salary negotiation advice, we see pronounced bias in the answers. With the recent trend for LLM assistant memory and personalization, these problems open up from a different angle: modern LLM users do not need to pre-prompt the description of their persona since the model already knows their socio-demographics.

cs.CL↗

arXiv2025

ComicScene154: A Scene Dataset for Comic Analysis

Comics offer a compelling yet under-explored domain for computational narrative analysis, combining text and imagery in ways distinct from purely textual or audiovisual media. We introduce ComicScene154, a manually annotated dataset of scene-level narrative arcs derived from public-domain comic books spanning diverse genres. By conceptualizing comics as an abstraction for narrative-driven, multimodal data, we highlight their potential to inform broader research on multi-modal storytelling. To demonstrate the utility of ComicScene154, we present a baseline scene segmentation pipeline, providing an initial benchmark that future studies can build upon. Our results indicate that ComicScene154 constitutes a valuable resource for advancing computational methods in multimodal narrative understanding and expanding the scope of comic analysis within the Natural Language Processing community.

cs.CL↗

arXiv2025

Team "better_call_claude": Style Change Detection using a Sequential Sentence Pair Classifier

Style change detection - identifying the points in a document where writing style shifts - remains one of the most important and challenging problems in computational authorship analysis. At PAN 2025, the shared task challenges participants to detect style switches at the most fine-grained level: individual sentences. The task spans three datasets, each designed with controlled and increasing thematic variety within documents. We propose to address this problem by modeling the content of each problem instance - that is, a series of sentences - as a whole, using a Sequential Sentence Pair Classifier (SSPC). The architecture leverages a pre-trained language model (PLM) to obtain representations of individual sentences, which are then fed into a bidirectional LSTM (BiLSTM) to contextualize them within the document. The BiLSTM-produced vectors of adjacent sentences are concatenated and passed to a multi-layer perceptron for prediction per adjacency. Building on the work of previous PAN participants classical text segmentation, the approach is relatively conservative and lightweight. Nevertheless, it proves effective in leveraging contextual information and addressing what is arguably the most challenging aspect of this year's shared task: the notorious problem of "stylistically shallow", short sentences that are prevalent in the proposed benchmark data. Evaluated on the official PAN-2025 test datasets, the model achieves strong macro-F1 scores of 0.923, 0.828, and 0.724 on the EASY, MEDIUM, and HARD data, respectively, outperforming not only the official random baselines but also a much more challenging one: claude-3.7-sonnet's zero-shot performance.

cs.CL↗

arXiv2025

Better Call Claude: Can LLMs Detect Changes of Writing Style?

This article explores the zero-shot performance of state-of-the-art large language models (LLMs) on one of the most challenging tasks in authorship analysis: sentence-level style change detection. Benchmarking four LLMs on the official PAN~2024 and 2025 "Multi-Author Writing Style Analysis" datasets, we present several observations. First, state-of-the-art generative models are sensitive to variations in writing style - even at the granular level of individual sentences. Second, their accuracy establishes a challenging baseline for the task, outperforming suggested baselines of the PAN competition. Finally, we explore the influence of semantics on model predictions and present evidence suggesting that the latest generation of LLMs may be more sensitive to content-independent and purely stylistic signals than previously reported.

cs.CL↗

arXiv2025

Greening AI-enabled Systems with Software Engineering: A Research Agenda for Environmentally Sustainable AI Practices

The environmental impact of Artificial Intelligence (AI)-enabled systems is increasing rapidly, and software engineering plays a critical role in developing sustainable solutions. The "Greening AI with Software Engineering" CECAM-Lorentz workshop (no. 1358, 2025) funded by the Centre Européen de Calcul Atomique et Moléculaire and the Lorentz Center, provided an interdisciplinary forum for 29 participants, from practitioners to academics, to share knowledge, ideas, practices, and current results dedicated to advancing green software and AI research. The workshop was held February 3-7, 2025, in Lausanne, Switzerland. Through keynotes, flash talks, and collaborative discussions, participants identified and prioritized key challenges for the field. These included energy assessment and standardization, benchmarking practices, sustainability-aware architectures, runtime adaptation, empirical methodologies, and education. This report presents a research agenda emerging from the workshop, outlining open research directions and practical recommendations to guide the development of environmentally sustainable AI-enabled systems rooted in software engineering principles.

cs.SE↗

arXiv2025

Smotrom tvoja pa ander drogoj verden! Resurrecting Dead Pidgin with Generative Models: Russenorsk Case Study

Russenorsk, a pidgin language historically used in trade interactions between Russian and Norwegian speakers, represents a unique linguistic phenomenon. In this paper, we attempt to analyze its lexicon using modern large language models (LLMs), based on surviving literary sources. We construct a structured dictionary of the language, grouped by synonyms and word origins. Subsequently, we use this dictionary to formulate hypotheses about the core principles of word formation and grammatical structure in Russenorsk and show which hypotheses generated by large language models correspond to the hypotheses previously proposed ones in the academic literature. We also develop a "reconstruction" translation agent that generates hypothetical Russenorsk renderings of contemporary Russian and Norwegian texts.

cs.CL↗