arXiv:2511.08877v1 Announce Type: new 
Abstract: Large language models (LLMs) have been increasingly applied to a wide range of tasks, from natural language understanding to code generation. While they have also been used to assist in citation recommendation, the hallucination of non-existent papers remains a major issue. Building on prior studies, this study hypothesizes that an LLM's ability to correctly produce bibliographic records depends on whether the underlying knowledge is generated or memorized, with highly cited papers (i.e., more frequently appear in the pretraining corpus) showing lower hallucination rates. We therefore assume citation count as a proxy for training data redundancy (i.e., the frequency with which a given bibliographic record appears in the pretraining corpus) and investigate how citation frequency affects hallucinated references in LLM outputs. Using GPT-4.1, we generated and manually verified 100 citations across twenty computer-science domains, and measured factual consistency via cosine similarity between generated and authentic metadata. The results revealed that (i) citation count is strongly correlated with factual accuracy, (ii) bibliographic information becomes almost verbatim memorized beyond roughly 1,000 citations, and (iii) memory interference occurs when multiple highly cited papers share similar content. These findings indicate a threshold where generalization shifts into memorization, with highly cited papers being nearly verbatim retained in the model.

تبحث دراسة جديدة نُشرت على arXiv العلاقة بين تكرار الاقتباسات ودقة السجلات الببليوغرافية التي تنتجها نماذج اللغة الكبيرة (LLMs). وتجد أن نماذج LLM، مثل GPT-4.1، تنتج اقتباسات أكثر دقة للأوراق ذات الاقتباسات العالية، مما يشير إلى أن النماذج تعتمد على المعلومات الم memorized بدلاً من توليدها. هذه الدراسة مهمة لأنها تتناول مشكلة الاقتباسات الوهمية في مخرجات نماذج LLM، والتي يمكن أن تقوض موثوقيتها في السياقات الأكاديمية.

Un nuevo estudio publicado en arXiv investiga la relación entre la frecuencia de citas y la precisión de los registros bibliográficos generados por modelos de lenguaje grandes (LLMs). Encuentra que los LLM, como GPT-4.1, producen citas más precisas para artículos muy citados, sugiriendo que los modelos dependen de información memorizada en lugar de generada. Esta investigación es significativa ya que aborda el problema de las referencias alucinadas en las salidas de los LLM, lo que puede socavar su fiabilidad en contextos académicos.

Une nouvelle étude publiée sur arXiv examine la relation entre la fréquence des citations et l'exactitude des références bibliographiques générées par les grands modèles de langage (LLM). Elle révèle que les LLM, comme GPT-4.1, produisent des citations plus précises pour les articles très cités, suggérant que les modèles s'appuient sur des informations mémorisées plutôt que générées. Cette recherche est significative car elle aborde le problème des références hallucinations dans les sorties des LLM, ce qui peut compromettre leur fiabilité dans les contextes académiques.

A new study published on arXiv investigates the relationship between citation frequency and the accuracy of bibliographic records generated by large language models (LLMs). It finds that LLMs, such as GPT-4.1, produce more accurate citations for highly cited papers, suggesting that the models rely on memorized information rather than generating it. This research is significant as it addresses the issue of hallucinated references in LLM outputs, which can undermine their reliability in academic contexts.

Hallucinate or Memorize? The Two Sides of Probabilistic Learning in Large Language Models

arXiv:2511.17481v1 Announce Type: new 
Abstract: World models learn to predict the temporal evolution of visual observations given a control signal, potentially enabling agents to reason about environments through forward simulation. Because of the focus on forward simulation, current world models generate predictions based on factual observations. For many emerging applications, such as comprehensive evaluations of physical AI behavior under varying conditions, the ability of world models to answer counterfactual queries, such as "what would happen if this object was removed?", is of increasing importance. We formalize counterfactual world models that additionally take interventions as explicit inputs, predicting temporal sequences under hypothetical modifications to observed scene properties. Traditional world models operate directly on entangled pixel-space representations where object properties and relationships cannot be selectively modified. This modeling choice prevents targeted interventions on specific scene properties. We introduce CWMDT, a framework to overcome those limitations, turning standard video diffusion models into effective counterfactual world models. First, CWMDT constructs digital twins of observed scenes to explicitly encode objects and their relationships, represented as structured text. Second, CWMDT applies large language models to reason over these representations and predict how a counterfactual intervention propagates through time to alter the observed scene. Third, CWMDT conditions a video diffusion model with the modified representation to generate counterfactual visual sequences. Evaluations on two benchmarks show that the CWMDT approach achieves state-of-the-art performance, suggesting that alternative representations of videos, such as the digital twins considered here, offer powerful control signals for video forward simulation-based world models.

تم تقديم إطار عمل جديد لنماذج العالم المضاد للحقائق، مما يسمح بتوقع تسلسلات زمنية تحت تعديلات افتراضية لخصائص المشهد المرصود. يعتمد هذا التقدم على نماذج العالم التقليدية التي تركز فقط على الملاحظات الواقعية، مما يمكّن من فهم أكثر دقة للبيئات من خلال المحاكاة المتقدمة.

Se ha introducido un nuevo marco para los modelos mundiales contrafactuales, que permite la predicción de secuencias temporales bajo modificaciones hipotéticas de las propiedades de la escena observada. Este avance se basa en los modelos mundiales tradicionales que se centran únicamente en observaciones fácticas, permitiendo una comprensión más matizada de los entornos a través de la simulación hacia adelante.

Un nouveau cadre pour les modèles mondiaux contrefactuels a été introduit, permettant la prédiction de séquences temporelles sous des modifications hypothétiques des propriétés de la scène observée. Cette avancée s'appuie sur les modèles mondiaux traditionnels qui se concentrent uniquement sur les observations factuelles, permettant une compréhension plus nuancée des environnements grâce à la simulation avancée.

A new framework for counterfactual world models has been introduced, which allows for the prediction of temporal sequences under hypothetical modifications to observed scene properties. This advancement builds on traditional world models that focus solely on factual observations, enabling a more nuanced understanding of environments through forward simulation.

Hallucinate or Memorize? The Two Sides of Probabilistic Learning in Large Language Models

Was this article worth reading? Share it

Sellm

Augmeta

Sourcely