Searcharxiv⌕ Search

arXiv subjects

Yadollah Yaghoobzadeh

Publications and source records attributed to Yadollah Yaghoobzadeh.

At least 19 recordsLinked to original sources
arXiv2026

CLIN: an Objective Framework for Evaluating Creativity in Short Persian Literary Text

Evaluating creativity in large language model (LLM) outputs remains challenging because creativity is multidimensional and human-centered. We examine how reliably LLMs evaluate short literary text in Persian, a low-resource language, across multiple evaluation strategies and prompt formulations. We find that LLM-human agreement varies substantially across dimensions: alignment is stronger for structured TTCT-derived properties such as Originality, Fluency, and Elaboration, but considerably weaker for more subjective dimensions, particularly Emotion and Attractiveness. Judgments are also sensitive to prompt formulation, while few-shot prompting, ensembling, and multi-agent debate provide no consistent improvement. Motivated by this dimension-dependent behavior, we investigate whether structured creativity dimensions can instead be approximated using simple, interpretable proxy metrics. We introduce CLIN, which evaluates three TTCT-derived dimensions separately using topic-aware novelty for Originality, contextual lexical clustering for Fluency, and lexical diversity for Elaboration. These proxies achieve human alignment comparable to or better than the strongest zero-shot LLM judge in our setting while requiring substantially lower evaluation cost.

cs.CL↗
arXiv2026

GhazalBench: Canonical Verse Access in LLMs across Persian Ghazals and Shakespearean Sonnets

Persian poetry plays an active role in Iranian cultural practice, where verses by canonical poets such as Hafez and Saadi are frequently quoted, paraphrased, or completed from incomplete cues. Supporting such interactions requires language models to reliably access canonical verses from semantic and lexical information. We introduce GhazalBench, a benchmark for evaluating how large language models (LLMs) access the canonical surface forms of Persian ghazal under usage-grounded conditions. Unlike prior work that primarily treats memorization as a liability, GhazalBench studies settings in which access to exact wording is functionally important. The benchmark evaluates completion and recognition under varied semantic and lexical cues. Across proprietary and open-weight multilingual LLMs, exact completion remains challenging, while recognition is substantially stronger. Performance also varies across poets and model families. Parallel experiments on Shakespearean sonnets yield markedly higher completion for several models, consistent with differences in exposure to canonical texts. {Additional experiments suggest that post-training may reduce direct continuation-based access to canonical text.} Our findings motivate evaluation frameworks that distinguish production from recognition and assess access to culturally significant canonical texts. GhazalBench is available at https://github.com/kalhorghazal/GhazalBench.

cs.CL↗
arXiv2026

Explanations of Large Language Models Explain Language Representations in the Brain

Large Language Model (LLM) representations are known to align with brain activity during language processing, but it remains unclear what drives this alignment. We test whether explainable AI (XAI) can help answer this: using attribution methods, we quantify the contribution of each input word to an LLM's next-word predictions and use these explanations to predict fMRI data from participants listening to narratives. We find that gradient-based attribution methods robustly align with brain activity, contribute unique variance beyond acoustic and word-rate confounds, and outperform internal representations in early auditory regions. Using conductance, we extend attribution from words to individual layers, asking what each layer's attribution reveals about the model's computation and how this relates to its brain alignment. Early layers show greater word-type sensitivity and align preferentially with auditory regions, whereas the final layer's attribution is dominated by positional information and exhibits broad cortical alignment. Together, these findings demonstrate that attribution-based explanations can be used not only to measure LLM--brain alignment but to characterize what it reflects.

cs.CL↗
arXiv2026

Layer-wise Positional Bias in Short-Context Language Modeling

Transformer language models systematically prefer tokens at specific input positions regardless of semantic relevance---a phenomenon known as positional bias. Prior work characterizes this bias in model behavior through performance drops in long-context tasks or in model architecture through attention-based analyses. However, it remains unmeasured how input positions actually drive predictions layer by layer. We introduce a layer conductance framework within a sliding-window design, applied to short-context next-word prediction to isolate model-internal behavior from task and context-window pressure. The resulting layer-wise positional importance profiles are stable across diverse texts and lexical scrambling, confirming they reflect model-internal structure. Characterizing how these profiles evolve across depth, we find recency bias increases monotonically while primacy bias is subtle and diminishes. We also find that this positional bias is not uniform across word types: function words exhibit higher recency bias while content words show higher primacy bias.

cs.CL↗
arXiv2026

Constrained Semantic Decompression in LLMs through Persian Proverb-Conditioned Story Generation

Transforming a dense, abstract proverb into an engaging and morally faithful narrative requires deep cultural understanding and robust semantic grounding. We frame this problem as a \emph{constrained semantic decompression} task and study proverb-conditioned story generation as a testbed for abstraction-to-realization in large language models (LLMs). Focusing on Persian, we introduce the Proverb Aligned Narrative Dataset (PAND), pairing proverbs with human-written stories and explicit meanings. By a hybrid evaluation framework that combines human-calibrated LLM-as-a-Judge with structural metrics, we analyze model behavior across multiple prompting regimes. Our findings reveal a persistent \emph{decompression gap}: current LLMs often achieve strong surface-level fluency while failing to faithfully instantiate the underlying moral and causal structure encoded in proverbs. We further show that explicit reasoning and iterative refinement can partially mitigate these failures, suggesting that many decompression errors arise from difficulties in translating abstract meaning into narrative form rather than a complete lack of relevant knowledge. Our proposed task naturally extends to other forms of compressed cultural knowledge.

cs.CL↗
arXiv2026

Synthia: Scalable Grounded Persona Generation from Social Media Data

Persona-driven simulations are increasingly used in computational social science, yet their validity critically depends on the fidelity of the underlying personas. Constructing virtual populations that are both authentic and scalable remains a central challenge. We introduce Synthia, a persona-generation framework that grounds LLM-generated personas in real social-media posts while delegating narrative construction to language models, using publicly available data from the Bluesky platform. Across multiple social-survey benchmarks, Synthia improves alignment with human opinion distributions over prior state-of-the-art approaches while relying on substantially smaller models. A multi-dimensional fairness and bias analysis shows that Synthia outperforms previous methods for most demographics across different dimensions. Uniquely, Synthia preserves interaction-graph structure among personas grounded in real social network users, enabling network-aware analysis, which we demonstrate through two homophily-focused case studies. Together, these results position Synthia as a practical and reliable framework for constructing scalable, high-fidelity, and equitable virtual populations.

cs.CL↗
arXiv2026

FFE-Hallu:Hallucinations in Fixed Figurative Expressions:Benchmark of Idioms and Proverbs in the Persian Language

Figurative language, particularly fixed figurative expressions (FFEs) such as idioms and proverbs, poses persistent challenges for large language models (LLMs). Unlike literal phrases, FFEs are culturally grounded, largely non-compositional, and conventionally fixed, making them especially vulnerable to figurative hallucination. We define figurative hallucination as the generation or endorsement of expressions that sound idiomatic and plausible but do not exist as authentic figurative expressions in the target language. We introduce FFEHallu, the first comprehensive benchmark for evaluating figurative hallucination in LLMs, with a focus on Persian, a linguistically rich yet underrepresented language. FFEHallu consists of 600 carefully curated instances spanning three complementary tasks: (i) FFE generation from meaning, (ii) detection of fabricated FFEs across four controlled construction categories, and (iii) FFE to FFE translation from English to Persian. Evaluating six state of the art multilingual LLMs, we find systematic weaknesses in figurative competence and cultural grounding. While models such as GPT4.1 demonstrate relatively strong performance in rejecting fabricated FFEs and retrieving authentic ones, most models struggle to reliably distinguish real expressions from high quality fabrications and frequently hallucinate during cross lingual translation. These findings reveal substantial gaps in current LLMs handling of figurative language and underscore the need for targeted benchmarks to assess and mitigate figurative hallucination.

cs.CL↗
arXiv2025

LLMCARE: early detection of cognitive impairment via transformer models enhanced by LLM-generated synthetic data

Alzheimer's disease and related dementias(ADRD) affect nearly five million older adults in the United States, yet more than half remain undiagnosed. Speech-based natural language processing(NLP) offers a scalable approach for detecting early cognitive decline through subtle linguistic markers that may precede clinical diagnosis. This study develops and evaluates a speech-based screening pipeline integrating transformer embeddings with handcrafted linguistic features, synthetic augmentation using large language models(LLMs), and benchmarking of unimodal and multimodal classifiers. External validation assessed generalizability to a MCI-only cohort. Transcripts were drawn from the ADReSSo 2021 benchmark dataset(n=237, Pitt Corpus) and the DementiaBank Delaware corpus(n=205, MCI vs. controls). Ten transformer models were tested under three fine-tuning strategies. A late-fusion model combined embeddings from the top transformer with 110 linguistic features. Five LLMs(LLaMA8B/70B, MedAlpaca7B, Ministral8B,GPT-4o) generated label-conditioned synthetic speech for augmentation, and three multimodal LLMs(GPT-4o,Qwen-Omni,Phi-4) were evaluated in zero-shot and fine-tuned modes. On ADReSSo, the fusion model achieved F1=83.3(AUC=89.5), outperforming transformer-only and linguistic baselines. MedAlpaca7B augmentation(2x) improved F1=85.7, though larger scales reduced gains. Fine-tuning boosted unimodal LLMs(MedAlpaca7B F1=47.7=>78.7), while multimodal models performed lower (Phi-4=71.6;GPT-4o=67.6). On Delaware, the fusion plus 1x MedAlpaca7B model achieved F1=72.8(AUC=69.6). Integrating transformer and linguistic features enhances ADRD detection. LLM-based augmentation improves data efficiency but yields diminishing returns, while current multimodal models remain limited. Validation on an independent MCI cohort supports the pipeline's potential for scalable, clinically relevant early screening.

cs.CL↗
arXiv2025

Evaluating the Creativity of LLMs in Persian Literary Text Generation

Large language models (LLMs) have demonstrated notable creative abilities in generating literary texts, including poetry and short stories. However, prior research has primarily centered on English, with limited exploration of non-English literary traditions and without standardized methods for assessing creativity. In this paper, we evaluate the capacity of LLMs to generate Persian literary text enriched with culturally relevant expressions. We build a dataset of user-generated Persian literary spanning 20 diverse topics and assess model outputs along four creativity dimensions-originality, fluency, flexibility, and elaboration-by adapting the Torrance Tests of Creative Thinking. To reduce evaluation costs, we adopt an LLM as a judge for automated scoring and validate its reliability against human judgments using intraclass correlation coefficients, observing strong agreement. In addition, we analyze the models' ability to understand and employ four core literary devices: simile, metaphor, hyperbole, and antithesis. Our results highlight both the strengths and limitations of LLMs in Persian literary text generation, underscoring the need for further refinement.

cs.CL↗
arXiv2025

A Dual-Axis Taxonomy of Knowledge Editing for LLMs: From Mechanisms to Functions

Large language models (LLMs) acquire vast knowledge from large text corpora, but this information can become outdated or inaccurate. Since retraining is computationally expensive, knowledge editing offers an efficient alternative -- modifying internal knowledge without full retraining. These methods aim to update facts precisely while preserving the model's overall capabilities. While existing surveys focus on the mechanism of editing (e.g., parameter changes vs. external memory), they often overlook the function of the knowledge being edited. This survey introduces a novel, complementary function-based taxonomy to provide a more holistic view. We examine how different mechanisms apply to various knowledge types -- factual, temporal, conceptual, commonsense, and social -- highlighting how editing effectiveness depends on the nature of the target knowledge. By organizing our review along these two axes, we map the current landscape, outline the strengths and limitations of existing methods, define the problem formally, survey evaluation tasks and datasets, and conclude with open challenges and future directions.

cs.AI↗
arXiv2025

Deconstructing the Crystal Ball: From Ad-Hoc Prediction to Principled Startup Evaluation with the SAISE Framework

The integration of Artificial Intelligence (AI) into startup evaluation represents a significant technological shift, yet the academic research underpinning this transition remains methodologically fragmented. Existing studies often employ ad-hoc approaches, leading to a body of work with inconsistent definitions of success, atheoretical features, and a lack of rigorous validation. This fragmentation severely limits the comparability, reliability, and practical utility of current predictive models. To address this critical gap, this paper presents a comprehensive systematic literature review of 57 empirical studies. We deconstruct the current state-of-the-art by systematically mapping the features, algorithms, data sources, and evaluation practices that define the AI-driven startup prediction landscape. Our synthesis reveals a field defined by a central paradox: a strong convergence on a common toolkit -- venture databases and tree-based ensembles -- but a stark divergence in methodological rigor. We identify four foundational weaknesses: a fragmented definition of "success," a divide between theory-informed and data-driven feature engineering, a chasm between common and best-practice model validation, and a nascent approach to data ethics and explainability. In response to these findings, our primary contribution is the proposal of the Systematic AI-driven Startup Evaluation (SAISE) Framework. This novel, five-stage prescriptive roadmap is designed to guide researchers from ad-hoc prediction toward principled evaluation. By mandating a coherent, end-to-end methodology that emphasizes stage-aware problem definition, theory-informed data synthesis, principled feature engineering, rigorous validation, and risk-aware interpretation, the SAISE framework provides a new standard for conducting more comparable, robust, and practically relevant research in this rapidly maturing domain

cs.CE↗
arXiv2025

GenKnowSub: Improving Modularity and Reusability of LLMs through General Knowledge Subtraction

Large language models often struggle with zero-shot generalization, and several modular approaches have been proposed to address this challenge. Yet, we hypothesize that a key limitation remains: the entanglement of general knowledge and task-specific adaptations. To overcome this, we propose a modular framework that disentangles these components by constructing a library of task-specific LoRA modules alongside a general-domain LoRA. By subtracting this general knowledge component from each task-specific module, we obtain residual modules that focus more exclusively on task-relevant information, a method we call general knowledge subtraction (GenKnowSub). Leveraging the refined task-specific modules and the Arrow routing algorithm \citep{ostapenko2024towards}, we dynamically select and combine modules for new inputs without additional training. Our studies on the Phi-3 model and standard Arrow as baselines reveal that using general knowledge LoRAs derived from diverse languages, including English, French, and German, yields consistent performance gains in both monolingual and cross-lingual settings across a wide set of benchmarks. Further experiments on Phi-2 demonstrate how GenKnowSub generalizes to weaker LLMs. The complete code and data are available at https://github.com/saharsamr/Modular-LLM.

cs.CL↗
arXiv2025

SOI Matters: Analyzing Multi-Setting Training Dynamics in Pretrained Language Models via Subsets of Interest

This work investigates the impact of multi-task, multi-lingual, and multi-source learning approaches on the robustness and performance of pretrained language models. To enhance this analysis, we introduce Subsets of Interest (SOI), a novel categorization framework that identifies six distinct learning behavior patterns during training, including forgettable examples, unlearned examples, and always correct examples. Through SOI transition heatmaps and dataset cartography visualization, we analyze how examples shift between these categories when transitioning from single-setting to multi-setting configurations. We perform comprehensive experiments across three parallel comparisons: multi-task vs. single-task learning using English tasks (entailment, paraphrase, sentiment), multi-source vs. single-source learning using sentiment analysis datasets, and multi-lingual vs. single-lingual learning using intent classification in French, English, and Persian. Our results demonstrate that multi-source learning consistently improves out-of-distribution performance by up to 7%, while multi-task learning shows mixed results with notable gains in similar task combinations. We further introduce a two-stage fine-tuning approach where the second stage leverages SOI-based subset selection to achieve additional performance improvements. These findings provide new insights into training dynamics and offer practical approaches for optimizing multi-setting language model performance.

cs.CL↗
arXiv2025

What Matters Most? A Quantitative Meta-Analysis of AI-Based Predictors for Startup Success

Background: Predicting startup success with machine learning is a rapidly growing field, yet findings on key predictors are often fragmented and context-specific. This makes it difficult to discern robust patterns and highlights a need for a systematic synthesis of the evidence. Methods: This study conducts a quantitative meta-analysis to synthesize the literature on predictor importance in AI-based startup evaluation. We performed a systematic review to identify a final sample of 13 empirical studies that report rankable feature importance. From these papers, we extracted and categorized 58 unique predictors, synthesizing their importance using a Weighted Importance Score (WIS) that balances a feature's average rank with its frequency of appearance. We also conducted a moderator analysis to investigate how predictor importance changes with context (e.g., success definition). Results: Our aggregate analysis reveals that the most consistently powerful predictors are a quartet of foundational attributes: Firm Characteristics (e.g., age, location), Investor Structure (e.g., investor quality), Digital and Social Traction (e.g., online momentum), and Funding History. The moderator analysis further reveals that this hierarchy is highly context-dependent. For instance, predicting near-term funding milestones elevates the importance of the deal's immediate context, while predicting long-term exits prioritizes fundamental firm and investor characteristics. Conclusion: The factors that best predict startup success are not universal but are contingent on the startup's goals, stage, and the data used for evaluation. Our findings point to a potential "convenience bias" in the literature, where predictor importance may be tied to data accessibility. We conclude by underscoring the need for standardized reporting practices to enable more robust, cumulative knowledge building in the field.

cs.CE↗
arXiv2025

Comparative Study of Multilingual Idioms and Similes in Large Language Models

This study addresses the gap in the literature concerning the comparative performance of LLMs in interpreting different types of figurative language across multiple languages. By evaluating LLMs using two multilingual datasets on simile and idiom interpretation, we explore the effectiveness of various prompt engineering strategies, including chain-of-thought, few-shot, and English translation prompts. We extend the language of these datasets to Persian as well by building two new evaluation sets. Our comprehensive assessment involves both closed-source (GPT-3.5, GPT-4o mini, Gemini 1.5), and open-source models (Llama 3.1, Qwen2), highlighting significant differences in performance across languages and figurative types. Our findings reveal that while prompt engineering methods are generally effective, their success varies by figurative type, language, and model. We also observe that open-source models struggle particularly with low-resource languages in similes. Additionally, idiom interpretation is nearing saturation for many languages, necessitating more challenging evaluations.

cs.CL↗
arXiv2025

Counterfactuals As a Means for Evaluating Faithfulness of Attribution Methods in Autoregressive Language Models

Despite the widespread adoption of autoregressive language models, explainability evaluation research has predominantly focused on span infilling and masked language models. Evaluating the faithfulness of an explanation method -- how accurately it explains the inner workings and decision-making of the model -- is challenging because it is difficult to separate the model from its explanation. Most faithfulness evaluation techniques corrupt or remove input tokens deemed important by a particular attribution (feature importance) method and observe the resulting change in the model's output. However, for autoregressive language models, this approach creates out-of-distribution inputs due to their next-token prediction training objective. In this study, we propose a technique that leverages counterfactual generation to evaluate the faithfulness of attribution methods for autoregressive language models. Our technique generates fluent, in-distribution counterfactuals, making the evaluation protocol more reliable.

cs.CL↗
arXiv2025

Large Language Models for Persian $ \leftrightarrow $ English Idiom Translation

Large language models (LLMs) have shown superior capabilities in translating figurative language compared to neural machine translation (NMT) systems. However, the impact of different prompting methods and LLM-NMT combinations on idiom translation has yet to be thoroughly investigated. This paper introduces two parallel datasets of sentences containing idiomatic expressions for Persian$\rightarrow$English and English$\rightarrow$Persian translations, with Persian idioms sampled from our PersianIdioms resource, a collection of 2,200 idioms and their meanings, with 700 including usage examples. Using these datasets, we evaluate various open- and closed-source LLMs, NMT models, and their combinations. Translation quality is assessed through idiom translation accuracy and fluency. We also find that automatic evaluation methods like LLM-as-a-judge, BLEU, and BERTScore are effective for comparing different aspects of model performance. Our experiments reveal that Claude-3.5-Sonnet delivers outstanding results in both translation directions. For English$\rightarrow$Persian, combining weaker LLMs with Google Translate improves results, while Persian$\rightarrow$English translations benefit from single prompts for simpler models and complex prompts for advanced ones.

cs.CL↗
arXiv2025

PerCul: A Story-Driven Cultural Evaluation of LLMs in Persian

Large language models predominantly reflect Western cultures, largely due to the dominance of English-centric training data. This imbalance presents a significant challenge, as LLMs are increasingly used across diverse contexts without adequate evaluation of their cultural competence in non-English languages, including Persian. To address this gap, we introduce PerCul, a carefully constructed dataset designed to assess the sensitivity of LLMs toward Persian culture. PerCul features story-based, multiple-choice questions that capture culturally nuanced scenarios. Unlike existing benchmarks, PerCul is curated with input from native Persian annotators to ensure authenticity and to prevent the use of translation as a shortcut. We evaluate several state-of-the-art multilingual and Persian-specific LLMs, establishing a foundation for future research in cross-cultural NLP evaluation. Our experiments demonstrate a 11.3% gap between best closed source model and layperson baseline while the gap increases to 21.3% by using the best open-weight model. You can access the dataset from here: https://huggingface.co/datasets/teias-ai/percul

cs.CL↗