Himanshu Beniwal

Himanshu Beniwal, Mayank Singh

Large language models frequently generate toxic, hateful, or harmful content, yet existing mitigation methods rely on costly retraining or output-level filtering with no mechanistic insight into where toxicity originates internally. We introduce Meow2X and TRNE, two complementary retraining-free frameworks that localize toxicity to specific layers and neurons by analyzing activation differentials between toxic and neutral prompts, then suppress them via inference-time scaling or minimal rank-one weight edits -- without any gradient descent. Evaluations across five LMs, two benchmarks, and 90 configurations using dual safety evaluators demonstrate consistent toxicity reduction while preserving language modeling quality. Our analysis reveals that toxicity is disproportionately encoded in early MLP layers, varies across architectures, and is systematically underestimated by single-evaluator setups -- underscoring the need for multi-evaluator safety assessment. By bridging mechanistic interpretability with practical detoxification, our framework offers a principled path toward safer, more transparent language models.

Manan Uppadhyay, Prashant Kodali, Pranjal Chitale et al.

Multilingual Large Language Models (mLLMs) leaderboards report per-language accuracy but rarely explain why disparities emerge, leaving systemic biases unattributed and offering practitioners no actionable levers. We first establish that these gaps are systematic rather than artifacts of sampling noise via distribution-free Friedman and Kruskal--Wallis tests, then introduce a two-step Bayesian hierarchical framework that decomposes multilingual performance variance into interpretable components. First, isolating the variance attributable to language identity, we show that observable language features (script, family, typological distance) explain $R^2_{\text{ling}} = 79\%$ of this variance on understanding tasks and $92\%$ on reasoning, with a model's internal representational similarity to English emerging as the dominant predictor across both task buckets. Second, decomposing the full (model$\times$benchmark$\times$language) cube, we find that NLU and reasoning have fundamentally divergent variance profiles: model identity dominates understanding ($66.7\%$ of variance), whereas the benchmark$\times$model interaction dominates reasoning ($46.3\%$). Together these results recast multilingual evaluation from passive performance mapping into an explainable, diagnostic framework with concrete levers for targeting the root drivers of language disparity.

Rajvee Sheth, Shubh Nisar, Heenaben Prajapati et al.

As the NLP community increasingly addresses challenges associated with multilingualism, robust annotation tools are essential to handle multilingual datasets efficiently. In this paper, we introduce a code-mixed multilingual text annotation framework, COMMENTATOR, specifically designed for annotating code-mixed text. The tool demonstrates its effectiveness in token-level and sentence-level language annotation tasks for Hinglish text. We perform robust qualitative human-based evaluations to showcase COMMENTATOR led to 5x faster annotations than the best baseline. Our code is publicly available at \url{https://github.com/lingo-iitgn/commentator}. The demonstration video is available at \url{https://bit.ly/commentator_video}.

Arya Shah, Himanshu beniwal, Mayank Singh

Aligning multilingual assistants with culturally grounded user preferences is essential for serving India's linguistically diverse population of over one billion speakers across multiple scripts. However, existing benchmarks either focus on a single language or conflate retrieval with generation, leaving open the question of whether current embedding models can encode persona-instruction compatibility without relying on response synthesis. We present a unified benchmark spanning 12 Indian languages and four evaluation tasks: monolingual and cross-lingual persona-to-instruction retrieval, reverse retrieval from instruction to persona, and binary compatibility classification. Eight multilingual embedding models are evaluated in a frozen-encoder setting with a thin logistic regression head for classification. E5-Large-Instruct achieves the highest Recall@1 of 27.4\% on monolingual retrieval and 20.7\% on cross-lingual transfer, while BGE-M3 leads reverse retrieval at 32.1\% Recall@1. For classification, LaBSE attains 75.3\% AUROC with strong calibration. These findings offer practical guidance for model selection in Indic multilingual retrieval and establish reproducible baselines for future work\footnote{Code, datasets, and models are publicly available at https://github.com/aryashah2k/PI-Indic-Align.

Himanshu Beniwal, Dishant Patel, Kowsik Nandagopan D et al.

Large Language Models (LLMs) are increasingly ubiquitous, yet their ability to retain and reason about temporal information remains limited, hindering their application in real-world scenarios where understanding the sequential nature of events is crucial. Our study experiments with 12 state-of-the-art models (ranging from 2B to 70B+ parameters) on a novel numerical-temporal dataset, \textbf{TempUN}, spanning from 10,000 BCE to 2100 CE, to uncover significant temporal retention and comprehension limitations. We propose six metrics to assess three learning paradigms to enhance temporal knowledge acquisition. Our findings reveal that open-source models exhibit knowledge gaps more frequently, suggesting a trade-off between limited knowledge and incorrect responses. Additionally, various fine-tuning approaches significantly improved performance, reducing incorrect outputs and impacting the identification of 'information not available' in the generations. The associated dataset and code are available at (https://github.com/lingoiitgn/TempUN).

Rajvee Sheth, Himanshu Beniwal, Mayank Singh

We introduce COMI-LINGUA, the largest manually annotated Hindi-English code-mixed dataset, comprising 125K+ high-quality instances across five core NLP tasks: Matrix Language Identification, Token-level Language Identification, Part-Of-Speech Tagging, Named Entity Recognition, and Machine Translation. Each instance is annotated by three bilingual annotators, yielding over 376K expert annotations with strong inter-annotator agreement (Fleiss' Kappa $\geq$ 0.81). The rigorously preprocessed and filtered dataset covers both Devanagari and Roman scripts and spans diverse domains, ensuring real-world linguistic coverage. Evaluation reveals that closed-source LLMs significantly outperform traditional tools and open-source models in zero-shot settings. Notably, one-shot prompting consistently boosts performance across tasks, especially in structure-sensitive predictions like POS and NER. Fine-tuning state-of-the-art LLMs on COMI-LINGUA demonstrates substantial improvements, achieving up to 95.25 F1 in NER, 98.77 F1 in MLI, and competitive MT performance, setting new benchmarks for Hinglish code-mixed text. COMI-LINGUA is publicly available at this URL: https://huggingface.co/datasets/LingoIITGN/COMI-LINGUA.

Rajvee Sheth, Samridhi Raj Sinha, Mahavir Patil et al.

Code-switching (CSW), the alternation of languages and scripts within a single utterance, remains a fundamental challenge for multilingual NLP, even amidst the rapid advances of large language models (LLMs). Most LLMs still struggle with mixed-language inputs, limited CSW datasets, and evaluation biases, hindering deployment in multilingual societies. This survey provides the first comprehensive analysis of CSW-aware LLM research, reviewing 308 studies spanning five research areas, 12 NLP tasks, 30+ datasets, and 80+ languages. We classify recent advances by architecture, training strategy, and evaluation methodology, outlining how LLMs have reshaped CSW modeling and what challenges persist. The paper concludes with a roadmap emphasizing the need for inclusive datasets, fair evaluation, and linguistically grounded models to achieve truly multilingual intelligence. A curated collection of all resources is maintained at https://github.com/lingo-iitgn/awesome-code-mixing/.

Himanshu Beniwal, Youngwoo Kim, Maarten Sap et al.

As large language models (LLMs) become increasingly prevalent in global applications, ensuring that they are toxicity-free across diverse linguistic contexts remains a critical challenge. We explore "Cross-lingual Detoxification", a cross-lingual paradigm that mitigates toxicity, enabling detoxification capabilities to transfer between high and low-resource languages across different script families. We analyze cross-lingual detoxification's effectiveness through 392 extensive settings to evaluate toxicity reduction in cross-distribution settings with limited data and investigate how mitigation impacts model performance on non-toxic tasks, revealing trade-offs between safety and knowledge preservation. Our code and dataset are publicly available at https://github.com/himanshubeniwal/Breaking-mBad.

Himanshu Beniwal, Sailesh Panda, Birudugadda Srivibhav et al.

We explore \textbf{C}ross-lingual \textbf{B}ackdoor \textbf{AT}tacks (X-BAT) in multilingual Large Language Models (mLLMs), revealing how backdoors inserted in one language can automatically transfer to others through shared embedding spaces. Using toxicity classification as a case study, we demonstrate that attackers can compromise multilingual systems by poisoning data in a single language, with rare and high-occurring tokens serving as specific, effective triggers. Our findings expose a critical vulnerability that influences the model's architecture, resulting in a concealed backdoor effect during the information flow. Our code and data are publicly available https://github.com/himanshubeniwal/X-BAT.

Himanshu Beniwal, Kowsik Nandagopan D, Mayank Singh

The training of large language models (LLMs) necessitates substantial data and computational resources, and updating outdated LLMs entails significant efforts and resources. While numerous model editing techniques (METs) have emerged to efficiently update model outputs without retraining, their effectiveness in multilingual LLMs, where knowledge is stored in diverse languages, remains an underexplored research area. This research paper introduces the cross-lingual model editing (\textbf{XME}) paradigm, wherein a fact is edited in one language, and the subsequent update propagation is observed across other languages. To investigate the XME paradigm, we conducted experiments using BLOOM, mBERT, and XLM-RoBERTa using the two writing scripts: \textit{Latin} (English, French, and Spanish) and \textit{Indic} (Hindi, Gujarati, and Bengali). The results reveal notable performance limitations of state-of-the-art METs under the XME setting, mainly when the languages involved belong to two distinct script families. These findings highlight the need for further research and development of XME techniques to address these challenges. For more comprehensive information, the dataset used in this research and the associated code are publicly available at the following URL\url{https://github.com/lingo-iitgn/XME}.

Ankit Yadav, Himanshu Beniwal, Mayank Singh

Driven by the surge in code generation using large language models (LLMs), numerous benchmarks have emerged to evaluate these LLMs capabilities. We conducted a large-scale human evaluation of HumanEval and MBPP, two popular benchmarks for Python code generation, analyzing their diversity and difficulty. Our findings unveil a critical bias towards a limited set of programming concepts, neglecting most of the other concepts entirely. Furthermore, we uncover a worrying prevalence of easy tasks, potentially inflating model performance estimations. To address these limitations, we propose a novel benchmark, PythonSaga, featuring 185 hand-crafted prompts on a balanced representation of 38 programming concepts across diverse difficulty levels. The robustness of our benchmark is demonstrated by the poor performance of existing Code-LLMs.

Priyanshu Kumar, Devansh Jain, Akhila Yerukola et al. · allen-ai, cmu

Truly multilingual safety moderation efforts for Large Language Models (LLMs) have been hindered by a narrow focus on a small set of languages (e.g., English, Chinese) as well as a limited scope of safety definition, resulting in significant gaps in moderation capabilities. To bridge these gaps, we release POLYGUARD, a new state-of-the-art multilingual safety model for safeguarding LLM generations, and the corresponding training and evaluation datasets. POLYGUARD is trained on POLYGUARDMIX, the largest multilingual safety training corpus to date containing 1.91M samples across 17 languages (e.g., Chinese, Czech, English, Hindi). We also introduce POLYGUARDPROMPTS, a high quality multilingual benchmark with 29K samples for the evaluation of safety guardrails. Created by combining naturally occurring multilingual human-LLM interactions and human-verified machine translations of an English-only safety dataset (WildGuardMix; Han et al., 2024), our datasets contain prompt-output pairs with labels of prompt harmfulness, response harmfulness, and response refusal. Through extensive evaluations across multiple safety and toxicity benchmarks, we demonstrate that POLYGUARD outperforms existing state-of-the-art open-weight and commercial safety classifiers by 5.5%. Our contributions advance efforts toward safer multilingual LLMs for all global users.

Himanshu Beniwal, Reddybathuni Venkat, Rohit Kumar et al.

This work introduces UnityAI-Guard, a framework for binary toxicity classification targeting low-resource Indian languages. While existing systems predominantly cater to high-resource languages, UnityAI-Guard addresses this critical gap by developing state-of-the-art models for identifying toxic content across diverse Brahmic/Indic scripts. Our approach achieves an impressive average F1-score of 84.23% across seven languages, leveraging a dataset of 567k training instances and 30k manually verified test instances. By advancing multilingual content moderation for linguistically diverse regions, UnityAI-Guard also provides public API access to foster broader adoption and application.

Youngwoo Kim, Himanshu Beniwal, Steven L. Johnson et al.

Effective content moderation systems require explicit classification criteria, yet online communities like subreddits often operate with diverse, implicit standards. This work introduces a novel approach to identify and extract these implicit criteria from historical moderation data using an interpretable architecture. We represent moderation criteria as score tables of lexical expressions associated with content removal, enabling systematic comparison across different communities. Our experiments demonstrate that these extracted lexical patterns effectively replicate the performance of neural moderation models while providing transparent insights into decision-making processes. The resulting criteria matrix reveals significant variations in how seemingly shared norms are actually enforced, uncovering previously undocumented moderation patterns including community-specific tolerances for language, features for topical restrictions, and underlying subcategories of the toxic speech classification.