arXiv subjects

Utkarsh Vaidya

Publications and source records attributed to Utkarsh Vaidya.

6 recordsLinked to original sources

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

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

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

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↗

arXiv2024

Adapting Multilingual LLMs to Low-Resource Languages using Continued Pre-training and Synthetic Corpus

Multilingual LLMs support a variety of languages; however, their performance is suboptimal for low-resource languages. In this work, we emphasize the importance of continued pre-training of multilingual LLMs and the use of translation-based synthetic pre-training corpora for improving LLMs in low-resource languages. We conduct our study in the context of the low-resource Indic language Hindi. We introduce Nemotron-Mini-Hindi 4B, a bilingual SLM supporting both Hindi and English, based on Nemotron-Mini 4B. The model is trained using a mix of real and synthetic Hindi + English tokens, with continuous pre-training performed on 400B tokens. We demonstrate that both the base and instruct models achieve state-of-the-art results on Hindi benchmarks while remaining competitive on English tasks. Additionally, we observe that the continued pre-training approach enhances the model's overall factual accuracy. We perform an ablation study to highlight the impact of Hindi pre-training, showing significant improvements in Hindi chat capabilities and factual accuracy, which cannot be achieved through Hindi alignment alone.

cs.CL↗