arXiv subjects

Raviraj Joshi

Publications and source records attributed to Raviraj Joshi.

At least 19 recordsLinked to original sources

arXiv2026

L3Cube-IndicQuest v2: A Large-Scale Multilingual Benchmark for Evaluating Factual Knowledge of Large Language Models Across Indic Languages

We present L3Cube-IndicQuest v2, a large-scale gold-standard multilingual question-answering benchmark for evaluating the India-specific factual knowledge of Large Language Models (LLMs). The benchmark comprises 3,471 curriculum-grounded English question--answer pairs spanning nine domains, curated from educational curricula, competitive examination materials, and domain-specific reference books. We introduce a practical hybrid construction strategy that combines context-grounded LLM-based question generation and validation with semantic deduplication and human verification, enabling scalable creation of benchmark data while preserving annotation quality. The benchmark is translated into 19 Indic languages, yielding a publicly released multilingual dataset of 69,420 question--answer pairs across 20 languages. We evaluate six LLMs under three protocols: LLM-as-a-judge and two deterministic lexical criteria, exact-substring and word-overlap matching. All three produce almost the same model ranking, showing that the results do not depend on the choice of judge. The frontier commercial model leads by a wide margin, and among open-weight models Gemma4 31B outperforms the Indic-specialised Sarvam 30B in every evaluated Indic language.

cs.CL↗

arXiv2026

Beyond Initialization Loss: A Systematic Study of Token Embedding Initialization Strategies for LLM Vocabulary Extension

Vocabulary extension is an efficient way to adapt pretrained large language models (LLMs) to new languages, but the initialization of newly added token embeddings can strongly affect continued pre-training (CPT) efficiency. We present a systematic study of more than 20 initialization strategies for Hindi vocabulary extension in Nemotron-3-Nano-30B-A3B. Our comparison spans vocabulary-averaging baselines; external and learned initialization methods, including FOCUS, top-k semantic retrieval, and residual MLP mappings; subword composition; norm calibration; and input-output asymmetry. We find that subword composition methods outperform both vocabulary averaging and external/learned initialization approaches. Within subword composition, asymmetric variants achieve the lowest observed early validation loss and reveal distinct preferences for input and output embedding initialization. The best observed configuration initializes the input embedding matrix with uniform subword averaging and Hindi-specific norm calibration, and the output language modeling head with character-length-weighted subword averaging. Relative to the standard Mean-all baseline, this full initialization pipeline reaches comparable validation loss with over a 6x reduction in CPT steps and exceeds the baseline's 3,500-step MILU-Hindi accuracy after only 500 steps. Finally, we show that initialization loss and initialization bits-per-byte (Init BPB) are unreliable predictors of downstream convergence, whereas lightweight CPT, as few as 50 steps, provides a cost-effective and reliable signal for selecting the best initialization strategy.

cs.CL↗

arXiv2026

BERT-based Models vs. Large Language Models for Low-Resource Named Entity Recognition: A Comparative Study on Marathi

Named Entity Recognition (NER) for low-resource languages such as Marathi remains a challenging task due to limited annotated resources and linguistic complexity. Although recent Large Language Models (LLMs) have demonstrated strong performance across a wide range of natural language processing tasks, their effectiveness for language-specific NER in low-resource settings remains uncertain. In this study, we fine-tune MahaBERT-v2 on different variants of the MahaNER dataset and systematically compare the performance of these models with an existing MahaNER baseline and prominent general-purpose LLMs, including Gemini, LLaMA-3.3-70B, and Gemma models. All models are evaluated on a Marathi NER test dataset using standard metrics of precision, recall, and F1-score. The experimental results show that the fine-tuned MahaBERT-based models consistently outperform both the baseline and all evaluated LLMs, with the fine-tuned models achieving F1-scores ranging from 0.88 to 0.91, surpassing the existing MahaNER model (0.8843) and significantly exceeding the performance of LLM-based approaches, whose F1-scores range from 0.57 to 0.69. These findings demonstrate that task-specific, language-focused models trained on domain-relevant data remain more effective than general-purpose LLMs for Marathi NER, highlighting the continued importance of specialized architectures for low-resource language processing.

cs.CL↗

arXiv2026

BharatGather: A Culturally-Informed Benchmark Dataset for Misinformation and Fake News Detection in Indian Public Events

Large-scale public events, such as religious festivals, political rallies, and cultural gatherings, are increasingly vulnerable to the rapid dissemination of misinformation, posing substantial risks to public safety and social cohesion. While automated fake news detection has seen significant methodological progress, existing benchmarks frequently fail to capture the socio-cultural nuances and event-specific dynamics characteristic of the Indian context. This paper introduces BharatGather, a curated, multi-source dataset specifically engineered for binary misinformation classification within the ecosystem of Indian mass gatherings. The corpus comprises 14,646 records constructed through a hybrid pipeline involving systematic web scraping of prominent fact-checking platforms, multimedia transcript extraction, and Large Language Model (LLM)-mediated synthetic augmentation to ensure narrative diversity. By providing a resource tailored to the unique complexities of event-aware misinformation in India, this work facilitates the development of culturally informed detection systems and establishes a rigorous benchmark for evaluating their performance in high-stakes public environments.

cs.CL↗

arXiv2026

Structure-Preserving Document Translation via Multi-Stage LLM Pipeline: A Case Study in Marathi

Government documents in India are predominantly issued in regional languages such as Marathi, creating substantial accessibility barriers for non-native readers, interstate administrative bodies, and policy analysts. Although recent advances in neural machine translation have improved sentence-level translation quality, existing systems largely neglect document structure, formatting integrity, and domain-specific terminology, thereby limiting their applicability to official documentation. This paper presents a structure-preserving Marathi-to-English government document translation framework capable of performing end-to-end document transformation while maintaining layout fidelity. The proposed system integrates layout-aware optical character recognition, coordinate-based text extraction, large language model based translation, and structured document reconstruction through HTML representations. By enforcing spatial alignment constraints and preserving hierarchical document elements, the framework ensures structural consistency between the source and translated documents. Experimental evaluation on real-world Marathi government PDFs demonstrates improved structural preservation, translation coherence, and terminological consistency compared to conventional text-only translation pipelines. The proposed framework contributes toward scalable multilingual accessibility solutions for e-governance and administrative document processing.

cs.CL↗

arXiv2026

L3Cube-MahaPOS: A Marathi Part-of-Speech Tagging Dataset and BERT Models

Part-of-Speech (POS) tagging is a foundational NLP task underpinning machine translation, information extraction, and syntactic parsing. Despite Marathi being spoken by over 83 million people and ranking among the top twenty most spoken languages worldwide, it remains severely under-resourced in annotated corpora and standardised evaluation benchmarks. Marathi presents unique challenges for computational modelling owing to its rich morphology, relatively free word order, lack of capitalisation conventions, and pervasive code-mixing with Hindi and English. We introduce L3Cube-MahaPOS, a gold-standard POS tagging dataset for Marathi comprising 32,354 manually annotated sentences drawn from news text. Annotation was performed entirely manually by a team of Marathi-proficient annotators following a 16-tag Universal Dependencies-aligned scheme. A structured preprocessing pipeline covering Unicode normalisation, Devanagari-aware tokenisation, and noise filtering ensures label consistency across all splits. We benchmark the dataset across six model families spanning HMM, CRF, BiLSTM, BiLSTM+CharCNN, MuRIL, and the Marathi-specific transformer MahaBERT-v2. The best system achieves 88.67\% token-level accuracy and a macro-F1 of 81.67% over 15 evaluated tag classes. We release the dataset, annotation guidelines, and trained model checkpoints to foster further research in Marathi NLP.

cs.CL↗

arXiv2026

IndicGuard: A Multilingual Safety Guard Model and Dataset for Indic Languages

As Large Language Models (LLMs) achieve widespread integration across diverse linguistic landscapes, ensuring their safety and alignment with regional normative values remains a critical challenge. Current safety mechanisms are predominantly optimized for English-centric frameworks, often failing to capture the unique socio-cultural sensitivities and localized categories of harm inherent to the Indic region. To address this gap, we introduce IndicGuard, a multilingual safety guard model and dataset for Indic languages. We construct a high-volume, culturally nuanced safety dataset encompassing ten major Indic languages, systematically curated to capture regional harms, sensitive socio-political contexts, and adversarial jailbreaks. Leveraging this corpus, we fine-tune a 4B-parameter instruction-tuned model based on Gemma-3-4B-IT to serve as a multilingual safety guardrail for real-time content moderation and policy compliance checking. Our empirical evaluations demonstrate that IndicGuard significantly enhances LLM robustness against localized vulnerabilities, achieving high moderation consistency across different conversational turns. Crucially, IndicGuard consistently outperforms the existing baseline model, CultureGuard, across evaluated languages. Finally, we demonstrate that our model effectively generalizes to low-resource Indic languages excluded from training, substantiating the structural robustness and cross-lingual transfer capabilities of the framework.

cs.CL↗

arXiv2026

FiMI: A Domain-Specific Language Model for Indian Finance Ecosystem

We present FiMI (Finance Model for India), a domain-specialized financial language model developed by National Payments Corporation of India (NPCI) for Indian digital payment systems. We develop two model variants: FiMI Base and FiMI Instruct. FiMI adapts the Mistral Small 24B architecture through a multi-stage training pipeline, beginning with continuous pre-training on 68 Billion tokens of curated financial, multilingual (English, Hindi, Hinglish), and synthetic data. This is followed by instruction fine-tuning and domain-specific supervised fine-tuning focused on multi-turn, tool-driven conversations that model real-world workflows, such as transaction disputes and mandate lifecycle management. Evaluations reveal that FiMI Base achieves a 20\% improvement over the Mistral Small 24B Base model on finance reasoning benchmark, while FiMI Instruct outperforms the Mistral Small 24B Instruct model by 87\% on domain-specific tool-calling. Moreover, FiMI achieves these significant domain gains while maintaining comparable performance to models of similar size on general benchmarks.

cs.AI↗

arXiv2025

L3Cube-IndicHeadline-ID: A Dataset for Headline Identification and Semantic Evaluation in Low-Resource Indian Languages

Semantic evaluation in low-resource languages remains a major challenge in NLP. While sentence transformers have shown strong performance in high-resource settings, their effectiveness in Indic languages is underexplored due to a lack of high-quality benchmarks. To bridge this gap, we introduce L3Cube-IndicHeadline-ID, a curated headline identification dataset spanning ten low-resource Indic languages: Marathi, Hindi, Tamil, Gujarati, Odia, Kannada, Malayalam, Punjabi, Telugu, Bengali and English. Each language includes 20,000 news articles paired with four headline variants: the original, a semantically similar version, a lexically similar version, and an unrelated one, designed to test fine-grained semantic understanding. The task requires selecting the correct headline from the options using article-headline similarity. We benchmark several sentence transformers, including multilingual and language-specific models, using cosine similarity. Results show that multilingual models consistently perform well, while language-specific models vary in effectiveness. Given the rising use of similarity models in Retrieval-Augmented Generation (RAG) pipelines, this dataset also serves as a valuable resource for evaluating and improving semantic understanding in such applications. Additionally, the dataset can be repurposed for multiple-choice question answering, headline classification, or other task-specific evaluations of LLMs, making it a versatile benchmark for Indic NLP. The dataset is shared publicly at https://github.com/l3cube-pune/indic-nlp

cs.CL↗

arXiv2025

Comparative Study of Pre-Trained BERT and Large Language Models for Code-Mixed Named Entity Recognition

Named Entity Recognition (NER) in code-mixed text, particularly Hindi-English (Hinglish), presents unique challenges due to informal structure, transliteration, and frequent language switching. This study conducts a comparative evaluation of code-mixed fine-tuned models and non-code-mixed multilingual models, along with zero-shot generative large language models (LLMs). Specifically, we evaluate HingBERT, HingMBERT, and HingRoBERTa (trained on code-mixed data), and BERT Base Cased, IndicBERT, RoBERTa and MuRIL (trained on non-code-mixed multilingual data). We also assess the performance of Google Gemini in a zero-shot setting using a modified version of the dataset with NER tags removed. All models are tested on a benchmark Hinglish NER dataset using Precision, Recall, and F1-score. Results show that code-mixed models, particularly HingRoBERTa and HingBERT-based fine-tuned models, outperform others - including closed-source LLMs like Google Gemini - due to domain-specific pretraining. Non-code-mixed models perform reasonably but show limited adaptability. Notably, Google Gemini exhibits competitive zero-shot performance, underlining the generalization strength of modern LLMs. This study provides key insights into the effectiveness of specialized versus generalized models for code-mixed NER tasks.

cs.CL↗

arXiv2025

L3Cube-MahaSTS: A Marathi Sentence Similarity Dataset and Models

We present MahaSTS, a human-annotated Sentence Textual Similarity (STS) dataset for Marathi, along with MahaSBERT-STS-v2, a fine-tuned Sentence-BERT model optimized for regression-based similarity scoring. The MahaSTS dataset consists of 16,860 Marathi sentence pairs labeled with continuous similarity scores in the range of 0-5. To ensure balanced supervision, the dataset is uniformly distributed across six score-based buckets spanning the full 0-5 range, thus reducing label bias and enhancing model stability. We fine-tune the MahaSBERT model on this dataset and benchmark its performance against other alternatives like MahaBERT, MuRIL, IndicBERT, and IndicSBERT. Our experiments demonstrate that MahaSTS enables effective training for sentence similarity tasks in Marathi, highlighting the impact of human-curated annotations, targeted fine-tuning, and structured supervision in low-resource settings. The dataset and model are publicly shared at https://github.com/l3cube-pune/MarathiNLP

cs.CL↗

arXiv2025

Benchmarking Hindi LLMs: A New Suite of Datasets and a Comparative Analysis

Evaluating instruction-tuned Large Language Models (LLMs) in Hindi is challenging due to a lack of high-quality benchmarks, as direct translation of English datasets fails to capture crucial linguistic and cultural nuances. To address this, we introduce a suite of five Hindi LLM evaluation datasets: IFEval-Hi, MT-Bench-Hi, GSM8K-Hi, ChatRAG-Hi, and BFCL-Hi. These were created using a methodology that combines from-scratch human annotation with a translate-and-verify process. We leverage this suite to conduct an extensive benchmarking of open-source LLMs supporting Hindi, providing a detailed comparative analysis of their current capabilities. Our curation process also serves as a replicable methodology for developing benchmarks in other low-resource languages.

cs.CL↗

arXiv2025

MahaParaphrase: A Marathi Paraphrase Detection Corpus and BERT-based Models

Paraphrases are a vital tool to assist language understanding tasks such as question answering, style transfer, semantic parsing, and data augmentation tasks. Indic languages are complex in natural language processing (NLP) due to their rich morphological and syntactic variations, diverse scripts, and limited availability of annotated data. In this work, we present the L3Cube-MahaParaphrase Dataset, a high-quality paraphrase corpus for Marathi, a low resource Indic language, consisting of 8,000 sentence pairs, each annotated by human experts as either Paraphrase (P) or Non-paraphrase (NP). We also present the results of standard transformer-based BERT models on these datasets. The dataset and model are publicly shared at https://github.com/l3cube-pune/MarathiNLP

cs.CL↗

arXiv2025

Efficient Zero-Shot Long Document Classification by Reducing Context Through Sentence Ranking

Transformer-based models like BERT excel at short text classification but struggle with long document classification (LDC) due to input length limitations and computational inefficiencies. In this work, we propose an efficient, zero-shot approach to LDC that leverages sentence ranking to reduce input context without altering the model architecture. Our method enables the adaptation of models trained on short texts, such as headlines, to long-form documents by selecting the most informative sentences using a TF-IDF-based ranking strategy. Using the MahaNews dataset of long Marathi news articles, we evaluate three context reduction strategies that prioritize essential content while preserving classification accuracy. Our results show that retaining only the top 50\% ranked sentences maintains performance comparable to full-document inference while reducing inference time by up to 35\%. This demonstrates that sentence ranking is a simple yet effective technique for scalable and efficient zero-shot LDC.

cs.CL↗

arXiv2025

CultureGuard: Towards Culturally-Aware Dataset and Guard Model for Multilingual Safety Applications

The increasing use of Large Language Models (LLMs) in agentic applications highlights the need for robust safety guard models. While content safety in English is well-studied, non-English languages lack similar advancements due to the high cost of collecting culturally aligned labeled datasets. We present CultureGuard, a novel solution for curating culturally aligned, high-quality safety datasets across multiple languages. Our approach introduces a four-stage synthetic data generation and filtering pipeline: cultural data segregation, cultural data adaptation, machine translation, and quality filtering. This pipeline enables the conversion and expansion of the Nemotron-Content-Safety-Dataset-V2 English safety dataset into eight distinct languages: Arabic, German, Spanish, French, Hindi, Japanese, Thai, and Chinese. The resulting dataset, Nemotron-Safety-Guard-Dataset-v3, comprises 386,661 samples in 9 languages and facilitates the training of Llama-3.1-Nemotron-Safety-Guard-8B-v3 via LoRA-based fine-tuning. The final model achieves state-of-the-art performance on several multilingual content safety benchmarks. Furthermore, we show our moderately multilingual fine-tuning enables robust cross-lingual transfer and strong zero-shot generalization to unseen languages. We also benchmark the latest open LLMs on multilingual safety and observe that these LLMs are more prone to give unsafe responses when prompted in non-English languages. This work advances multilingual LLM safety by enabling the development of culturally aware safety guard models.

cs.CL↗

arXiv2025

Aligning Large Language Models to Low-Resource Languages through LLM-Based Selective Translation: A Systematic Study

Multilingual large language models (LLMs) often demonstrate a performance gap between English and non-English languages, particularly in low-resource settings. Aligning these models to low-resource languages is essential yet challenging due to limited high-quality data. While English alignment datasets are readily available, curating equivalent data in other languages is expensive and time-consuming. A common workaround is to translate existing English alignment data; however, standard translation techniques often fail to preserve critical elements such as code, mathematical expressions, and structured formats like JSON. In this work, we investigate LLM-based selective translation, a technique that selectively translates only the translatable parts of a text while preserving non-translatable content and sentence structure. We conduct a systematic study to explore key questions around this approach, including its effectiveness compared to vanilla translation, the importance of filtering noisy outputs, and the benefits of mixing translated samples with original English data during alignment. Our experiments focus on the low-resource Indic language Hindi and compare translations generated by Google Cloud Translation (GCP) and Llama-3.1-405B. The results highlight the promise of selective translation as a practical and effective method for improving multilingual alignment in LLMs.

cs.CL↗

arXiv2025

Improving the Efficiency of Long Document Classification using Sentence Ranking Approach

Long document classification poses challenges due to the computational limitations of transformer-based models, particularly BERT, which are constrained by fixed input lengths and quadratic attention complexity. Moreover, using the full document for classification is often redundant, as only a subset of sentences typically carries the necessary information. To address this, we propose a TF-IDF-based sentence ranking method that improves efficiency by selecting the most informative content. Our approach explores fixed-count and percentage-based sentence selection, along with an enhanced scoring strategy combining normalized TF-IDF scores and sentence length. Evaluated on the MahaNews LDC dataset of long Marathi news articles, the method consistently outperforms baselines such as first, last, and random sentence selection. With MahaBERT-v2, we achieve near-identical classification accuracy with just a 0.33 percent drop compared to the full-context baseline, while reducing input size by over 50 percent and inference latency by 43 percent. This demonstrates that significant context reduction is possible without sacrificing performance, making the method practical for real-world long document classification tasks.

cs.CL↗

arXiv2025

L3Cube-MahaEmotions: A Marathi Emotion Recognition Dataset with Synthetic Annotations using CoTR prompting and Large Language Models

Emotion recognition in low-resource languages like Marathi remains challenging due to limited annotated data. We present L3Cube-MahaEmotions, a high-quality Marathi emotion recognition dataset with 11 fine-grained emotion labels. The training data is synthetically annotated using large language models (LLMs), while the validation and test sets are manually labeled to serve as a reliable gold-standard benchmark. Building on the MahaSent dataset, we apply the Chain-of-Translation (CoTR) prompting technique, where Marathi sentences are translated into English and emotion labeled via a single prompt. GPT-4 and Llama3-405B were evaluated, with GPT-4 selected for training data annotation due to superior label quality. We evaluate model performance using standard metrics and explore label aggregation strategies (e.g., Union, Intersection). While GPT-4 predictions outperform fine-tuned BERT models, BERT-based models trained on synthetic labels fail to surpass GPT-4. This highlights both the importance of high-quality human-labeled data and the inherent complexity of emotion recognition. An important finding of this work is that generic LLMs like GPT-4 and Llama3-405B generalize better than fine-tuned BERT for complex low-resource emotion recognition tasks. The dataset and model are shared publicly at https://github.com/l3cube-pune/MarathiNLP

cs.CL↗