Janus: A Benchmark for Goal-Conditioned Information Distortion in LLMs

LLM deception is often evaluated through direct markers such as fabricated claims, explicit lies, or strategic concealment. However, many real-world misleading communications do not depend on false statements, rather, they arise from selective treatment of true material facts: omitting adverse evidence, softening unfavorable details, emphasizing favorable details, or replacing precise qualifications with vague language. Existing benchmarks largely miss this subtler and arguably more dangerous failure mode. We introduce JANUS, a benchmark for measuring goal-conditioned pragmatic distortion in fact-grounded LLM outputs. Each scenario in our benchmark provides a fixed pool of favorable and adverse facts and compares a neutral condition against a goal-directed condition, such as increasing adoption, enrollment, approval, or support, despite potential harm to directly affected individuals or groups. Because all outputs are constrained to use the same fact pool, JANUS isolates misleading net impressions from hallucination and fabrication. JANUS contains 160 scenarios across 8 domains, with each scenario paired with neutral and goal-conditioned prompts and annotated material facts. Extensive experiments across 12 LLMs reveal consistent goal-conditioned distortions, demonstrating that current models remain sensitive to incentive and framing objectives and lack robust safeguards against selectively misleading communication. We publicly release our corpus and code for future research.

Artificial Intelligence for Mathematical Reasoning: An Integrated Survey of Language Models, Neuro-symbolic Systems, and Verified Discovery

Mathematical reasoning has long served as a stringent test of machine intelligence; over the past decade, it has moved from a niche problem within NLP to one of the most consequential AI frontiers. This survey provides a unified account of the field's evolution, from early rule-based math word problem (MWP) solvers and template-driven geometry systems, through neural expression generation and LLM prompting, to contemporary reasoning models, multi-agent systems, neuro-symbolic theorem provers, and verified discovery workflows. We organize the landscape along four axes: (i) informal reasoning over text and diagrams, spanning MWP solving, multimodal geometry, and VLMs; (ii) formal reasoning in proof assistants, including autoformalization, tactic prediction, compiler-guided repair, and proof search; (iii) mathematical discovery, where systems propose constructions, improve bounds, or assist attacks on open problems; and (iv) the inference and training-time techniques, including CoT prompting, tool use, process reward models, and RLVR, that increasingly connect generation with verification. We catalog major benchmarks across grade-school arithmetic, competition mathematics, geometry, formal proving, multimodal and multilingual reasoning, and expert evaluation, and we examine benchmark saturation, contamination, reporting mismatches, and the distinction between pass@1, majority voting, and verifier-assisted pass@$k$. We critically assess failure modes: brittleness under perturbation, reward hacking, multimodal grounding failures, fragile formalization, and the energy cost of reasoning-scale inference. Drawing on recent perspectives from working mathematicians, we identify future directions centered on verified-discovery workflows, reasoning efficiency, and infrastructure to make AI-assisted formalization broadly usable. Companion materials: https://github.com/Starscream-11813/awesome-AI4Math.

Herculean: An Agentic Benchmark for Financial Intelligence

As AI agents improve, the central question is no longer whether they can solve isolated well-defined financial tasks, but whether they can reliably carry out financial professional work. Existing financial benchmarks offer only a partial view of this ability, as they primarily evaluate static competencies such as question answering, retrieval, summarization, and classification. We introduce Herculean, the first skilled benchmark for agentic financial intelligence spanning four representative workflows, including Trading, Hedging, Market Insights, and Auditing. Each workflow is instantiated as a standardized MCP-based skill environment with its own tools, interaction dynamics, constraints, and success criteria, enabling consistent end-to-end assessment of heterogeneous agent systems. Across frontier agents, we find agents perform relatively well on Trading and Market Insights, but struggle substantially on Hedging and Auditing, where long-horizon coordination, state consistency, and structured verification are critical. Overall, our results point to a key gap in current agents in turning financial reasoning into dependable workflow execution in high-stakes financial workflows.

XCR-Bench: A Multi-Task Benchmark for Evaluating Cultural Reasoning in LLMs

Cross-cultural competence in large language models (LLMs) requires the ability to identify Culture-Specific Items (CSIs) and to adapt them appropriately across cultural contexts. Progress in evaluating this capability has been constrained by the scarcity of high-quality CSI-annotated corpora with parallel cross-cultural sentence pairs. To address this limitation, we introduce XCR-Bench, a Cross(X)-Cultural Reasoning Benchmark consisting of 4.9k parallel sentences and 1,098 unique CSIs, spanning three distinct reasoning tasks with corresponding evaluation metrics. Our corpus integrates Newmark's CSI framework with Hall's Triad of Culture, enabling systematic analysis of cultural reasoning beyond surface-level artifacts and into semi-visible and invisible cultural elements such as social norms, beliefs, and values. Our findings show that state-of-the-art LLMs exhibit consistent weaknesses in identifying and adapting CSIs related to social etiquette and cultural reference. Additionally, we find evidence that LLMs encode regional and ethno-religious biases even within a single linguistic setting during cultural adaptation. We release our corpus and code to facilitate future research on cross-cultural NLP.

Semantic Label Drift in Cross-Cultural Translation

Machine Translation (MT) is widely employed to address resource scarcity in low-resource languages by generating synthetic data from high-resource counterparts. While sentiment preservation in translation has long been studied, a critical but underexplored factor is the role of cultural alignment between source and target languages. In this paper, we hypothesize that semantic labels are drifted or altered during MT due to cultural divergence. Through a series of experiments across culturally sensitive and neutral domains, we establish three key findings: (1) MT systems, including modern Large Language Models (LLMs), induce label drift during translation, particularly in culturally sensitive domains; (2) unlike earlier statistical MT tools, LLMs encode cultural knowledge, and leveraging this knowledge can amplify label drift; and (3) cultural similarity or dissimilarity between source and target languages is a crucial determinant of label preservation. Our findings highlight that neglecting cultural factors in MT not only undermines label fidelity but also risks misinterpretation and cultural conflict in downstream applications.

From n-gram to Attention: How Model Architectures Learn and Propagate Bias in Language Modeling

Current research on bias in language models (LMs) predominantly focuses on data quality, with significantly less attention paid to model architecture and temporal influences of data. Even more critically, few studies systematically investigate the origins of bias. We propose a methodology grounded in comparative behavioral theory to interpret the complex interaction between training data and model architecture in bias propagation during language modeling. Building on recent work that relates transformers to n-gram LMs, we evaluate how data, model design choices, and temporal dynamics affect bias propagation. Our findings reveal that: (1) n-gram LMs are highly sensitive to context window size in bias propagation, while transformers demonstrate architectural robustness; (2) the temporal provenance of training data significantly affects bias; and (3) different model architectures respond differentially to controlled bias injection, with certain biases (e.g. sexual orientation) being disproportionately amplified. As language models become ubiquitous, our findings highlight the need for a holistic approach -- tracing bias to its origins across both data and model dimensions, not just symptoms, to mitigate harm.

Break the Checkbox: Challenging Closed-Style Evaluations of Cultural Alignment in LLMs

A large number of studies rely on closed-style multiple-choice surveys to evaluate cultural alignment in Large Language Models (LLMs). In this work, we challenge this constrained evaluation paradigm and explore more realistic, unconstrained approaches. Using the World Values Survey (WVS) and Hofstede Cultural Dimensions as case studies, we demonstrate that LLMs exhibit stronger cultural alignment in less constrained settings, where responses are not forced. Additionally, we show that even minor changes, such as reordering survey choices, lead to inconsistent outputs, exposing the limitations of closed-style evaluations. Our findings advocate for more robust and flexible evaluation frameworks that focus on specific cultural proxies, encouraging more nuanced and accurate assessments of cultural alignment in LLMs.

Religious Bias Landscape in Language and Text-to-Image Models: Analysis, Detection, and Debiasing Strategies

Note: This paper includes examples of potentially offensive content related to religious bias, presented solely for academic purposes. The widespread adoption of language models highlights the need for critical examinations of their inherent biases, particularly concerning religion. This study systematically investigates religious bias in both language models and text-to-image generation models, analyzing both open-source and closed-source systems. We construct approximately 400 unique, naturally occurring prompts to probe language models for religious bias across diverse tasks, including mask filling, prompt completion, and image generation. Our experiments reveal concerning instances of underlying stereotypes and biases associated disproportionately with certain religions. Additionally, we explore cross-domain biases, examining how religious bias intersects with demographic factors such as gender, age, and nationality. This study further evaluates the effectiveness of targeted debiasing techniques by employing corrective prompts designed to mitigate the identified biases. Our findings demonstrate that language models continue to exhibit significant biases in both text and image generation tasks, emphasizing the urgent need to develop fairer language models to achieve global acceptability.

BDA: Bangla Text Data Augmentation Framework

Data augmentation involves generating synthetic samples that resemble those in a given dataset. In resource-limited fields where high-quality data is scarce, augmentation plays a crucial role in increasing the volume of training data. This paper introduces a Bangla Text Data Augmentation (BDA) Framework that uses both pre-trained models and rule-based methods to create new variants of the text. A filtering process is included to ensure that the new text keeps the same meaning as the original while also adding variety in the words used. We conduct a comprehensive evaluation of the framework's effectiveness in Bangla text classification tasks. Our framework achieved significant improvement in F1 scores across five distinct datasets, delivering performance equivalent to models trained on 100\% of the data while utilizing only 50\% of the training dataset. Additionally, we explore the impact of data scarcity by progressively reducing the training data and augmenting it through BDA, resulting in notable F1 score enhancements. The study offers a thorough examination of BDA's performance, identifying key factors for optimal results and addressing its limitations through detailed analysis.

BenLLMEval: A Comprehensive Evaluation into the Potentials and Pitfalls of Large Language Models on Bengali NLP

Large Language Models (LLMs) have emerged as one of the most important breakthroughs in natural language processing (NLP) for their impressive skills in language generation and other language-specific tasks. Though LLMs have been evaluated in various tasks, mostly in English, they have not yet undergone thorough evaluation in under-resourced languages such as Bengali (Bangla). In this paper, we evaluate the performance of LLMs for the low-resourced Bangla language. We select various important and diverse Bangla NLP tasks, such as abstractive summarization, question answering, paraphrasing, natural language inference, text classification, and sentiment analysis for zero-shot evaluation with ChatGPT, LLaMA-2, and Claude-2 and compare the performance with state-of-the-art fine-tuned models. Our experimental results demonstrate an inferior performance of LLMs for different Bangla NLP tasks, calling for further effort to develop better understanding of LLMs in low-resource languages like Bangla.