Mayank Singh

BigScience Workshop, Teven Le Scao, Angela Fan et al. · allen-ai, berkeley

Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions. While these capabilities have led to widespread adoption, most LLMs are developed by resource-rich organizations and are frequently kept from the public. As a step towards democratizing this powerful technology, we present BLOOM, a 176B-parameter open-access language model designed and built thanks to a collaboration of hundreds of researchers. BLOOM is a decoder-only Transformer language model that was trained on the ROOTS corpus, a dataset comprising hundreds of sources in 46 natural and 13 programming languages (59 in total). We find that BLOOM achieves competitive performance on a wide variety of benchmarks, with stronger results after undergoing multitask prompted finetuning. To facilitate future research and applications using LLMs, we publicly release our models and code under the Responsible AI License.

Pritam Kadasi, Anuj Tiwari, Mayank Singh

Our submission presented in this paper is for SemEval-2026 Task 9: Multilingual Text Classification Challenge - Polarization Detection and it covers all three subtasks: (1) binary polarization detection, (2) polarization type classification and (3) polarization manifestation identification. We adopt a systematic approach of research on short designed prompts by considering twelve designed prompts that are different in terminology clarity, detail of the definition, guidance of reasoning and in-context examples use. The experiments are conducted using aya-101 and Gemma3-27B, with the latter chosen for the submission at the end of the development through performance considerations. Our system has an average macro level F1-score of 0.762 on Subtask 1, 0.587 on Subtask 2 and 0.444 on Subtask 3 with the average accuracy of 0.819, 0.678 and 0.498, respectively, on the official test set averaged among 22 languages, respectively. With cross-task and cross-lingual analysis, we demonstrate that prompt-based approaches can be used effectively to detect coarse grained polarization but encounter more and more difficulties as far as fine-grained and multi-label sociolinguistic classification is concerned.

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.

Dwip Dalal, Vivek Srivastava, Mayank Singh

Social media plays a significant role in cross-cultural communication. A vast amount of this occurs in code-mixed and multilingual form, posing a significant challenge to Natural Language Processing (NLP) tools for processing such information, like language identification, topic modeling, and named-entity recognition. To address this, we introduce a large-scale multilingual, and multi-topic dataset (MMT) collected from Twitter (1.7 million Tweets), encompassing 13 coarse-grained and 63 fine-grained topics in the Indian context. We further annotate a subset of 5,346 tweets from the MMT dataset with various Indian languages and their code-mixed counterparts. Also, we demonstrate that the currently existing tools fail to capture the linguistic diversity in MMT on two downstream tasks, i.e., topic modeling and language identification. To facilitate future research, we have make the anonymized and annotated dataset available at https://huggingface.co/datasets/LingoIITGN/MMT.

Arya Shah, Vaibhav Tripathi, Mayank Singh et al.

Vision-language models are increasingly deployed in high-stakes settings, yet their susceptibility to sycophantic manipulation remains poorly understood, particularly in relation to how these models represent visual information internally. Whether models whose visual representations more closely mirror human neural processing are also more resistant to adversarial pressure is an open question with implications for both neuroscience and AI safety. We investigate this question by evaluating 12 open-weight vision-language models spanning 6 architecture families and a 40$\times$ parameter range (256M--10B) along two axes: brain alignment, measured by predicting fMRI responses from the Natural Scenes Dataset across 8 human subjects and 6 visual cortex regions of interest, and sycophancy, measured through 76,800 two-turn gaslighting prompts spanning 5 categories and 10 difficulty levels. Region-of-interest analysis reveals that alignment specifically in early visual cortex (V1--V3) is a reliable negative predictor of sycophancy ($r = -0.441$, BCa 95\% CI $[-0.740, -0.031]$), with all 12 leave-one-out correlations negative and the strongest effect for existence denial attacks ($r = -0.597$, $p = 0.040$). This anatomically specific relationship is absent in higher-order category-selective regions, suggesting that faithful low-level visual encoding provides a measurable anchor against adversarial linguistic override in vision-language models. We release our code on \href{https://github.com/aryashah2k/Gaslight-Gatekeep-Sycophantic-Manipulation}{GitHub} and dataset on \href{https://huggingface.co/datasets/aryashah00/Gaslight-Gatekeep-V1-V3}{Hugging Face}

Rahul Gupta, Vivek Srivastava, Mayank Singh · amazon-science

The multi-sentential long sequence textual data unfolds several interesting research directions pertaining to natural language processing and generation. Though we observe several high-quality long-sequence datasets for English and other monolingual languages, there is no significant effort in building such resources for code-mixed languages such as Hinglish (code-mixing of Hindi-English). In this paper, we propose a novel task of identifying multi-sentential code-mixed text (MCT) from multilingual articles. As a use case, we leverage multilingual articles from two different data sources and build a first-of-its-kind multi-sentential code-mixed Hinglish dataset i.e., MUTANT. We propose a token-level language-aware pipeline and extend the existing metrics measuring the degree of code-mixing to a multi-sentential framework and automatically identify MCT in the multilingual articles. The MUTANT dataset comprises 67k articles with 85k identified Hinglish MCTs. To facilitate future research, we make the publicly available.

Nikolaos Gkanatsios, Mayank Singh, Zhaoyuan Fang et al.

We present Analogical Networks, a model that encodes domain knowledge explicitly, in a collection of structured labelled 3D scenes, in addition to implicitly, as model parameters, and segments 3D object scenes with analogical reasoning: instead of mapping a scene to part segments directly, our model first retrieves related scenes from memory and their corresponding part structures, and then predicts analogous part structures for the input scene, via an end-to-end learnable modulation mechanism. By conditioning on more than one retrieved memories, compositions of structures are predicted, that mix and match parts across the retrieved memories. One-shot, few-shot or many-shot learning are treated uniformly in Analogical Networks, by conditioning on the appropriate set of memories, whether taken from a single, few or many memory exemplars, and inferring analogous parses. We show Analogical Networks are competitive with state-of-the-art 3D segmentation transformers in many-shot settings, and outperform them, as well as existing paradigms of meta-learning and few-shot learning, in few-shot settings. Analogical Networks successfully segment instances of novel object categories simply by expanding their memory, without any weight updates. Our code and models are publicly available in the project webpage: http://analogicalnets.github.io/.

Pratik Kayal, Mrinal Anand, Harsh Desai et al.

Scientific documents contain tables that list important information in a concise fashion. Structure and content extraction from tables embedded within PDF research documents is a very challenging task due to the existence of visual features like spanning cells and content features like mathematical symbols and equations. Most existing table structure identification methods tend to ignore these academic writing features. In this paper, we adapt the transformer-based language modeling paradigm for scientific table structure and content extraction. Specifically, the proposed model converts a tabular image to its corresponding LaTeX source code. Overall, we outperform the current state-of-the-art baselines and achieve an exact match accuracy of 70.35 and 49.69% on table structure and content extraction, respectively. Further analysis demonstrates that the proposed models efficiently identify the number of rows and columns, the alphanumeric characters, the LaTeX tokens, and symbols.

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.

Rima Hazra, Mayank Singh, Pawan Goyal et al.

Interdisciplinarity has over the recent years have gained tremendous importance and has become one of the key ways of doing cutting edge research. In this paper we attempt to model the citation flow across three different fields -- Physics (PHY), Mathematics (MA) and Computer Science (CS). For instance, is there a specific pattern in which these fields cite one another? We carry out experiments on a dataset comprising more than 1.2 million articles taken from these three fields. We quantify the citation interactions among these three fields through temporal bucket signatures. We present numerical models based on variants of the recently proposed relay-linking framework to explain the citation dynamics across the three disciplines. These models make a modest attempt to unfold the underlying principles of how citation links could have been formed across the three fields over time.

Shruti Singh, Hitesh Lodwal, Husain Malwat et al.

Language models (LMs) are no longer restricted to ML community, and instruction-tuned LMs have led to a rise in autonomous AI agents. As the accessibility of LMs grows, it is imperative that an understanding of their capabilities, intended usage, and development cycle also improves. Model cards are a popular practice for documenting detailed information about an ML model. To automate model card generation, we introduce a dataset of 500 question-answer pairs for 25 ML models that cover crucial aspects of the model, such as its training configurations, datasets, biases, architecture details, and training resources. We employ annotators to extract the answers from the original paper. Further, we explore the capabilities of LMs in generating model cards by answering questions. Our initial experiments with ChatGPT-3.5, LLaMa, and Galactica showcase a significant gap in the understanding of research papers by these aforementioned LMs as well as generating factual textual responses. We posit that our dataset can be used to train models to automate the generation of model cards from paper text and reduce human effort in the model card curation process. The complete dataset is available on https://osf.io/hqt7p/?view_only=3b9114e3904c4443bcd9f5c270158d37

Shruti Singh, Mayank Singh

Language models are increasingly becoming popular in AI-powered scientific IR systems. This paper evaluates popular scientific language models in handling (i) short-query texts and (ii) textual neighbors. Our experiments showcase the inability to retrieve relevant documents for a short-query text even under the most relaxed conditions. Additionally, we leverage textual neighbors, generated by small perturbations to the original text, to demonstrate that not all perturbations lead to close neighbors in the embedding space. Further, an exhaustive categorization yields several classes of orthographically and semantically related, partially related, and completely unrelated neighbors. Retrieval performance turns out to be more influenced by the surface form rather than the semantics of the text.

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.

Shruti Singh, Shoaib Alam, Husain Malwat et al.

The ever-increasing volume of paper submissions makes it difficult to stay informed about the latest state-of-the-art research. To address this challenge, we introduce LEGOBench, a benchmark for evaluating systems that generate scientific leaderboards. LEGOBench is curated from 22 years of preprint submission data on arXiv and more than 11k machine learning leaderboards on the PapersWithCode portal. We present four graph-based and two language model-based leaderboard generation task configurations. We evaluate popular encoder-only scientific language models as well as decoder-only large language models across these task configurations. State-of-the-art models showcase significant performance gaps in automatic leaderboard generation on LEGOBench. The code is available on GitHub ( https://github.com/lingo-iitgn/LEGOBench ) and the dataset is hosted on OSF ( https://osf.io/9v2py/?view_only=6f91b0b510df498ba01595f8f278f94c ).

Mayank Singh, Vikas Yadav, Shiva Krishna Reddy Malay et al.

Automatic search for Multi-Agent Systems has recently emerged as a key focus in agentic AI research. Several prior approaches have relied on LLM-based free-form search over the code space. In this work, we propose a more structured framework that explores the same space through a fixed set of simple, composable components. We show that, despite lacking the generative flexibility of LLMs during the candidate generation stage, our method outperforms prior approaches on four out of five benchmarks across two domains: mathematics and question answering. Furthermore, our method offers additional advantages, including a more cost-efficient search process and the generation of modular, interpretable multi-agent systems with simpler logic.

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/.

Samridhi Raj Sinha, Rajvee Sheth, Abhishek Upperwal et al.

The rapid advancement of Large Language Models (LLMs) has intensified the need for evaluation frameworks that address the requirements of linguistically diverse regions, such as India, and go beyond English-centric benchmarks. We introduce EKA-EVAL, a unified evaluation framework that integrates over 35+ benchmarks (including 10 Indic benchmarks) across nine major evaluation categories. The framework provides broader coverage than existing Indian language evaluation tools, offering 11 core capabilities through a modular architecture, seamless integration with Hugging Face and proprietary models, and plug-and-play usability. As the first end-to-end suite for scalable, multilingual LLM benchmarking, the framework combines extensive benchmarks, modular workflows, and dedicated support for low-resource Indian languages to enable inclusive assessment of LLM capabilities across diverse domains. We conducted extensive comparisons against five existing baselines, demonstrating that EKA-EVAL achieves the highest participant ratings in four out of five categories. The framework is open-source and publicly available at: https://github.com/lingo-iitgn/eka-eval.

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}.

Shruti Singh, Mayank Singh, Pawan Goyal

Comparing research papers is a conventional method to demonstrate progress in experimental research. We present COMPARE, a taxonomy and a dataset of comparison discussions in peer reviews of research papers in the domain of experimental deep learning. From a thorough observation of a large set of review sentences, we build a taxonomy of categories in comparison discussions and present a detailed annotation scheme to analyze this. Overall, we annotate 117 reviews covering 1,800 sentences. We experiment with various methods to identify comparison sentences in peer reviews and report a maximum F1 Score of 0.49. We also pretrain two language models specifically on ML, NLP, and CV paper abstracts and reviews to learn informative representations of peer reviews. The annotated dataset and the pretrained models are available at https://github.com/shruti-singh/COMPARE .

Naman Jain, Ankush Chauhan, Atharva Chewale et al.

Song lyrics convey a meaningful story in a creative manner with complex rhythmic patterns. Researchers have been successful in generating and analyisng lyrics for poetry and songs in English and Chinese. But there are no works which explore the Hindi language datasets. Given the popularity of Hindi songs across the world and the ambiguous nature of romanized Hindi script, we propose Bollyrics, an automatic lyric generator for romanized Hindi songs. We propose simple techniques to capture rhyming patterns before and during the model training process in Hindi language. The dataset and codes are available publicly at https://github.com/lingo-iitgn/Bollyrics.

Mayank Singh, Barnopriyo Barua, Priyank Palod et al.

This paper proposes OCR++, an open-source framework designed for a variety of information extraction tasks from scholarly articles including metadata (title, author names, affiliation and e-mail), structure (section headings and body text, table and figure headings, URLs and footnotes) and bibliography (citation instances and references). We analyze a diverse set of scientific articles written in English language to understand generic writing patterns and formulate rules to develop this hybrid framework. Extensive evaluations show that the proposed framework outperforms the existing state-of-the-art tools with huge margin in structural information extraction along with improved performance in metadata and bibliography extraction tasks, both in terms of accuracy (around 50% improvement) and processing time (around 52% improvement). A user experience study conducted with the help of 30 researchers reveals that the researchers found this system to be very helpful. As an additional objective, we discuss two novel use cases including automatically extracting links to public datasets from the proceedings, which would further accelerate the advancement in digital libraries. The result of the framework can be exported as a whole into structured TEI-encoded documents. Our framework is accessible online at http://cnergres.iitkgp.ac.in/OCR++/home/.

Mayank Singh, Nazia Nafis, Abhijeet Kumar et al.

Global sustainable fund universe encompasses open-end funds and exchange-traded funds (ETF) that, by prospectus or other regulatory filings, claim to focus on Environment, Social and Governance (ESG). Challengingly, the claims can only be confirmed by examining the textual disclosures to check if there is presence of intentionality and ESG focus on its investment strategy. Currently, there is no regulation to enforce sustainability in ESG products space. This paper proposes a unique method and system to classify and score the fund prospectuses in the sustainable universe regarding specificity and transparency of language. We aim to employ few-shot learners to identify specific, ambiguous, and generic sustainable investment-related language. Additionally, we construct a ratio metric to determine language score and rating to rank products and quantify sustainability claims for US sustainable universe. As a by-product, we publish manually annotated quality training dataset on Hugging Face (ESG-Prospectus-Clarity-Category under cc-by-nc-sa-4.0) of more than 1K ESG textual statements. The performance of the few-shot finetuning approach is compared with zero-shot models e.g., Llama-13B, GPT 3.5 Turbo etc. We found that prompting large language models are not accurate for domain specific tasks due to misalignment issues. The few-shot finetuning techniques outperform zero-shot models by large margins of more than absolute ~30% in precision, recall and F1 metrics on completely unseen ESG languages (test set). Overall, the paper attempts to establish a systematic and scalable approach to measure and rate sustainability intention quantitatively for sustainable funds using texts in prospectus. Regulatory bodies, investors, and advisors may utilize the findings of this research to reduce cognitive load in investigating or screening of ESG funds which accurately reflects the ESG intention.

Pritam Kadasi, Mayank Singh

The NLP community has long advocated for the construction of multi-annotator datasets to better capture the nuances of language interpretation, subjectivity, and ambiguity. This paper conducts a retrospective study to show how performance scores can vary when a dataset expands from a single annotation per instance to multiple annotations. We propose a novel multi-annotator simulation process to generate datasets with varying annotation budgets. We show that similar datasets with the same annotation budget can lead to varying performance gains. Our findings challenge the popular belief that models trained on multi-annotation examples always lead to better performance than models trained on single or few-annotation examples.

Naman Jain, Pranjali Jain, Pratik Kayal et al.

India is an agro-based economy and proper information about agricultural practices is the key to optimal agricultural growth and output. In order to answer the queries of the farmer, we have build an agricultural chatbot based on the dataset from Kisan Call Center. This system is robust enough to answer queries related to weather, market rates, plant protection and government schemes. This system is available 24* 7, can be accessed through any electronic device and the information is delivered with the ease of understanding. The system is based on a sentence embedding model which gives an accuracy of 56%. After eliminating synonyms and incorporating entity extraction, the accuracy jumps to 86%. With such a system, farmers can progress towards easier information about farming related practices and hence a better agricultural output. The job of the Call Center workforce would be made easier and the hard work of various such workers can be redirected to a better goal.

Pritam Kadasi, Abhishek Upperwal, Mayank SIngh

We propose ADAPT, a meta-learning algorithm that \emph{learns} task sampling proportions under an explicit token budget for multi-task instruction tuning. Instead of fixing task weights by hand, \adapt{} maintains a continuous distribution over tasks and updates it via meta-gradients of a smooth worst-case validation objective, inducing an adaptive curriculum that allocates more tokens to useful tasks while avoiding collapse. We instantiate ADAPT on three $\sim$1B-parameter open-weight LLMs (Gemma-3-1B, LLaMA-3.2-1B, Qwen-0.6B), training on 20 Natural Instructions task types under budgets of $1\%$, $5\%$, and $10\%$ of the available supervised tokens, and compare against strong supervised fine-tuning baselines with uniform and size-proportional mixing. We conduct evaluations on 11 out-of-domain benchmarks spanning reasoning, reading comprehension, code generation, and instruction following, we find that ADAPT matches or slightly improves average downstream performance relative to the best static mixture, while using fewer effective training tokens and reallocating budget toward harder, benchmark-aligned tasks.

Pritam Kadasi, Abhishek Upperwal, Mayank Singh

Instruction tuning is now the default way to train and adapt large language models, but many instruction--input--output pairs are only weakly specified: for a given input, the same output can remain plausible under several alternative instructions. This raises a simple question: \emph{does the instruction uniquely determine the target output?} We propose the \textbf{Task--Specificity Score (TSS)} to quantify how much an instruction matters for predicting its output, by contrasting the true instruction against plausible alternatives for the same input. We further introduce \textbf{TSS++}, which uses hard alternatives and a small quality term to mitigate easy-negative effects. Across three instruction datasets (\textsc{Alpaca}, \textsc{Dolly-15k}, \textsc{NI-20}) and three open LLMs (Gemma, Llama, Qwen), we show that selecting task-specific examples improves downstream performance under tight token budgets and complements quality-based filters such as perplexity and IFD.

Sailesh Panda, Pritam Kadasi, Abhishek Upperwal et al.

Large language models (LLMs) often achieve strong performance on reasoning benchmarks, but final-answer accuracy alone does not show whether they faithfully execute the procedure specified in a prompt. We study this question through a controlled diagnostic benchmark for procedural execution, where models are given a step-wise arithmetic algorithm and two numeric inputs, and must return the final computed value. The benchmark uses simple arithmetic operations but increases complexity through algorithm length and look-back dependencies over intermediate variables. Across 14 models and 55 datasets, average first-answer accuracy drops from 61% on 5-step procedures to 20% on 95-step procedures. Generation-level analysis shows that failures often involve missing answers, premature answers, self-correction after an initial error, under-executed traces, and hallucinated extra steps. These findings suggest that apparent reasoning ability can mask substantial weaknesses in faithful instruction execution.

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.

Mayank Singh, Abhijeet Kumar, Sasidhar Donaparthi et al.

Data catalogs serve as repositories for organizing and accessing diverse collection of data assets, but their effectiveness hinges on the ease with which business users can look-up relevant content. Unfortunately, many data catalogs within organizations suffer from limited searchability due to inadequate metadata like asset descriptions. Hence, there is a need of content generation solution to enrich and curate metadata in a scalable way. This paper explores the challenges associated with metadata creation and proposes a unique prompt enrichment idea of leveraging existing metadata content using retrieval based few-shot technique tied with generative large language models (LLM). The literature also considers finetuning an LLM on existing content and studies the behavior of few-shot pretrained LLM (Llama, GPT3.5) vis-à-vis few-shot finetuned LLM (Llama2-7b) by evaluating their performance based on accuracy, factual grounding, and toxicity. Our preliminary results exhibit more than 80% Rouge-1 F1 for the generated content. This implied 87%- 88% of instances accepted as is or curated with minor edits by data stewards. By automatically generating descriptions for tables and columns in most accurate way, the research attempts to provide an overall framework for enterprises to effectively scale metadata curation and enrich its data catalog thereby vastly improving the data catalog searchability and overall usability.

Pritam Kadasi, Sriman Reddy Kondam, Srivathsa Vamsi Chaturvedula et al.

With the massive surge in ML models on platforms like Hugging Face, users often lose track and struggle to choose the best model for their downstream tasks, frequently relying on model popularity indicated by download counts, likes, or recency. We investigate whether this popularity aligns with actual model performance and how the comprehensiveness of model documentation correlates with both popularity and performance. In our study, we evaluated a comprehensive set of 500 Sentiment Analysis models on Hugging Face. This evaluation involved massive annotation efforts, with human annotators completing nearly 80,000 annotations, alongside extensive model training and evaluation. Our findings reveal that model popularity does not necessarily correlate with performance. Additionally, we identify critical inconsistencies in model card reporting: approximately 80% of the models analyzed lack detailed information about the model, training, and evaluation processes. Furthermore, about 88% of model authors overstate their models' performance in the model cards. Based on our findings, we provide a checklist of guidelines for users to choose good models for downstream tasks.

Akash Ghosh, B Venkata Sahith, Niloy Ganguly et al.

Question-answering (QA) on hybrid scientific tabular and textual data deals with scientific information, and relies on complex numerical reasoning. In recent years, while tabular QA has seen rapid progress, understanding their robustness on scientific information is lacking due to absence of any benchmark dataset. To investigate the robustness of the existing state-of-the-art QA models on scientific hybrid tabular data, we propose a new dataset, "SciTabQA", consisting of 822 question-answer pairs from scientific tables and their descriptions. With the help of this dataset, we assess the state-of-the-art Tabular QA models based on their ability (i) to use heterogeneous information requiring both structured data (table) and unstructured data (text) and (ii) to perform complex scientific reasoning tasks. In essence, we check the capability of the models to interpret scientific tables and text. Our experiments show that "SciTabQA" is an innovative dataset to study question-answering over scientific heterogeneous data. We benchmark three state-of-the-art Tabular QA models, and find that the best F1 score is only 0.462.

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.

Mayank Singh, Vikas Yadav, Eduardo Blanco

Automatic Prompt Optimization (APO) is a powerful approach for extracting performance from large language models without modifying their weights. Many existing methods rely on trial-and-error, testing different prompts or in-context examples until a good configuration emerges, often consuming substantial compute. Recently, natural language feedback derived from execution logs has shown promise as a way to identify how prompts can be improved. However, most prior approaches operate in a bottom-up manner, iteratively adjusting the prompt based on feedback from individual problems, which can cause them to lose the global perspective. In this work, we propose Error Taxonomy-Guided Prompt Optimization (ETGPO), a prompt optimization algorithm that adopts a top-down approach. ETGPO focuses on the global failure landscape by collecting model errors, categorizing them into a taxonomy, and augmenting the prompt with guidance targeting the most frequent failure modes. Across multiple benchmarks spanning mathematics, question answering, and logical reasoning, ETGPO achieves accuracy that is comparable to or better than state-of-the-art methods, while requiring roughly one third of the optimization-phase token usage and evaluation budget.

Vivek Srivastava, Mayank Singh

In this shared task, we seek the participating teams to investigate the factors influencing the quality of the code-mixed text generation systems. We synthetically generate code-mixed Hinglish sentences using two distinct approaches and employ human annotators to rate the generation quality. We propose two subtasks, quality rating prediction and annotators' disagreement prediction of the synthetic Hinglish dataset. The proposed subtasks will put forward the reasoning and explanation of the factors influencing the quality and human perception of the code-mixed text.

Ayush Garg, Sammed S Kagi, Vivek Srivastava et al.

Code-mixing is a phenomenon of mixing words and phrases from two or more languages in a single utterance of speech and text. Due to the high linguistic diversity, code-mixing presents several challenges in evaluating standard natural language generation (NLG) tasks. Various widely popular metrics perform poorly with the code-mixed NLG tasks. To address this challenge, we present a metric independent evaluation pipeline MIPE that significantly improves the correlation between evaluation metrics and human judgments on the generated code-mixed text. As a use case, we demonstrate the performance of MIPE on the machine-generated Hinglish (code-mixing of Hindi and English languages) sentences from the HinGE corpus. We can extend the proposed evaluation strategy to other code-mixed language pairs, NLG tasks, and evaluation metrics with minimal to no effort.

Vivek Srivastava, Mayank Singh

Text generation is a highly active area of research in the computational linguistic community. The evaluation of the generated text is a challenging task and multiple theories and metrics have been proposed over the years. Unfortunately, text generation and evaluation are relatively understudied due to the scarcity of high-quality resources in code-mixed languages where the words and phrases from multiple languages are mixed in a single utterance of text and speech. To address this challenge, we present a corpus (HinGE) for a widely popular code-mixed language Hinglish (code-mixing of Hindi and English languages). HinGE has Hinglish sentences generated by humans as well as two rule-based algorithms corresponding to the parallel Hindi-English sentences. In addition, we demonstrate the inefficacy of widely-used evaluation metrics on the code-mixed data. The HinGE dataset will facilitate the progress of natural language generation research in code-mixed languages.

Mrinal Anand, Pratik Kayal, Mayank Singh

Automatic code synthesis from natural language descriptions is a challenging task. We witness massive progress in developing code generation systems for domain-specific languages (DSLs) employing sequence-to-sequence deep learning techniques in the recent past. In this paper, we specifically experiment with \textsc{AlgoLisp} DSL-based generative models and showcase the existence of significant dataset bias through different classes of adversarial examples. We also experiment with two variants of Transformer-based models that outperform all existing \textsc{AlgoLisp} DSL-based code generation baselines. Consistent with the current state-of-the-art systems, our proposed models, too, achieve poor performance under adversarial settings. Therefore, we propose several dataset augmentation techniques to reduce bias and showcase their efficacy using robust experimentation.

Viraj Shah, Shruti Singh, Mayank Singh

We present TweeNLP, a one-stop portal that organizes Twitter's natural language processing (NLP) data and builds a visualization and exploration platform. It curates 19,395 tweets (as of April 2021) from various NLP conferences and general NLP discussions. It supports multiple features such as TweetExplorer to explore tweets by topics, visualize insights from Twitter activity throughout the organization cycle of conferences, discover popular research papers and researchers. It also builds a timeline of conference and workshop submission deadlines. We envision TweeNLP to function as a collective memory unit for the NLP community by integrating the tweets pertaining to research papers with the NLPExplorer scientific literature search engine. The current system is hosted at http://nlpexplorer.org/twitter/CFP .

Vivek Srivastava, Mayank Singh

Code-mixing is a frequent communication style among multilingual speakers where they mix words and phrases from two different languages in the same utterance of text or speech. Identifying and filtering code-mixed text is a challenging task due to its co-existence with monolingual and noisy text. Over the years, several code-mixing metrics have been extensively used to identify and validate code-mixed text quality. This paper demonstrates several inherent limitations of code-mixing metrics with examples from the already existing datasets that are popularly used across various experiments.

Vivek Srivastava, Mayank Singh

Multilingualism refers to the high degree of proficiency in two or more languages in the written and oral communication modes. It often results in language mixing, a.k.a. code-mixing, when a multilingual speaker switches between multiple languages in a single utterance of a text or speech. This paper discusses the current state of the NLP research, limitations, and foreseeable pitfalls in addressing five real-world applications for social good crisis management, healthcare, political campaigning, fake news, and hate speech for multilingual societies. We also propose futuristic datasets, models, and tools that can significantly advance the current research in multilingual NLP applications for the societal good. As a representative example, we consider English-Hindi code-mixing but draw similar inferences for other language pairs

Pratik Kayal, Mrinal Anand, Harsh Desai et al.

Tables present important information concisely in many scientific documents. Visual features like mathematical symbols, equations, and spanning cells make structure and content extraction from tables embedded in research documents difficult. This paper discusses the dataset, tasks, participants' methods, and results of the ICDAR 2021 Competition on Scientific Table Image Recognition to LaTeX. Specifically, the task of the competition is to convert a tabular image to its corresponding LaTeX source code. We proposed two subtasks. In Subtask 1, we ask the participants to reconstruct the LaTeX structure code from an image. In Subtask 2, we ask the participants to reconstruct the LaTeX content code from an image. This report describes the datasets and ground truth specification, details the performance evaluation metrics used, presents the final results, and summarizes the participating methods. Submission by team VCGroup got the highest Exact Match accuracy score of 74% for Subtask 1 and 55% for Subtask 2, beating previous baselines by 5% and 12%, respectively. Although improvements can still be made to the recognition capabilities of models, this competition contributes to the development of fully automated table recognition systems by challenging practitioners to solve problems under specific constraints and sharing their approaches; the platform will remain available for post-challenge submissions at https://competitions.codalab.org/competitions/26979 .

Akhilesh Ravi, Amit Yadav, Jainish Chauhan et al.

The increasing use of dialogue agents makes it extremely desirable for them to understand and acknowledge the implied emotions to respond like humans with empathy. Chatbots using traditional techniques analyze emotions based on the context and meaning of the text and lack the understanding of emotions expressed through face. Emojis representing facial expressions present a promising way to express emotions. However, none of the AI systems utilizes emojis for empathetic conversation generation. We propose, SentEmojiBot, based on the SentEmoji dataset, to generate empathetic conversations with a combination of emojis and text. Evaluation metrics show that the BERT-based model outperforms the vanilla transformer model. A user study indicates that the dialogues generated by our model were understandable and adding emojis improved empathetic traits in conversations by 9.8%

Harsh Desai, Pratik Kayal, Mayank Singh

Information Extraction (IE) from the tables present in scientific articles is challenging due to complicated tabular representations and complex embedded text. This paper presents TabLeX, a large-scale benchmark dataset comprising table images generated from scientific articles. TabLeX consists of two subsets, one for table structure extraction and the other for table content extraction. Each table image is accompanied by its corresponding LATEX source code. To facilitate the development of robust table IE tools, TabLeX contains images in different aspect ratios and in a variety of fonts. Our analysis sheds light on the shortcomings of current state-of-the-art table extraction models and shows that they fail on even simple table images. Towards the end, we experiment with a transformer-based existing baseline to report performance scores. In contrast to the static benchmarks, we plan to augment this dataset with more complex and diverse tables at regular intervals.

Puneet Mangla, Nupur Kumari, Mayank Singh et al.

Recent advances in generative adversarial networks (GANs) have shown remarkable progress in generating high-quality images. However, this gain in performance depends on the availability of a large amount of training data. In limited data regimes, training typically diverges, and therefore the generated samples are of low quality and lack diversity. Previous works have addressed training in low data setting by leveraging transfer learning and data augmentation techniques. We propose a novel transfer learning method for GANs in the limited data domain by leveraging informative data prior derived from self-supervised/supervised pre-trained networks trained on a diverse source domain. We perform experiments on several standard vision datasets using various GAN architectures (BigGAN, SNGAN, StyleGAN2) to demonstrate that the proposed method effectively transfers knowledge to domains with few target images, outperforming existing state-of-the-art techniques in terms of image quality and diversity. We also show the utility of data instance prior in large-scale unconditional image generation.

Heer Ambavi, Kavita Vaishnaw, Udit Vyas et al.

The entire world is engulfed in the fight against the COVID-19 pandemic, leading to a significant surge in research experiments, government policies, and social media discussions. A multi-modal information access and data visualization platform can play a critical role in supporting research aimed at understanding and developing preventive measures for the pandemic. In this paper, we present a multi-faceted AI-based search and visualization engine, CovidExplorer. Our system aims to help researchers understand current state-of-the-art COVID-19 research, identify research articles relevant to their domain, and visualize real-time trends and statistics of COVID-19 cases. In contrast to other existing systems, CovidExplorer also brings in India-specific topical discussions on social media to study different aspects of COVID-19. The system, demo video, and the datasets are available at http://covidexplorer.in.

Parth Patel, Nupur Kumari, Mayank Singh et al.

Generative Adversarial Networks (GANs) coupled with self-supervised tasks have shown promising results in unconditional and semi-supervised image generation. We propose a self-supervised approach (LT-GAN) to improve the generation quality and diversity of images by estimating the GAN-induced transformation (i.e. transformation induced in the generated images by perturbing the latent space of generator). Specifically, given two pairs of images where each pair comprises of a generated image and its transformed version, the self-supervision task aims to identify whether the latent transformation applied in the given pair is same to that of the other pair. Hence, this auxiliary loss encourages the generator to produce images that are distinguishable by the auxiliary network, which in turn promotes the synthesis of semantically consistent images with respect to latent transformations. We show the efficacy of this pretext task by improving the image generation quality in terms of FID on state-of-the-art models for both conditional and unconditional settings on CIFAR-10, CelebA-HQ and ImageNet datasets. Moreover, we empirically show that LT-GAN helps in improving controlled image editing for CelebA-HQ and ImageNet over baseline models. We experimentally demonstrate that our proposed LT self-supervision task can be effectively combined with other state-of-the-art training techniques for added benefits. Consequently, we show that our approach achieves the new state-of-the-art FID score of 9.8 on conditional CIFAR-10 image generation.

Nidhin Harilal, Udit Bhatia, Mayank Singh

Projection of changes in extreme indices of climate variables such as temperature and precipitation are critical to assess the potential impacts of climate change on human-made and natural systems, including critical infrastructures and ecosystems. While impact assessment and adaptation planning rely on high-resolution projections (typically in the order of a few kilometers), state-of-the-art Earth System Models (ESMs) are available at spatial resolutions of few hundreds of kilometers. Current solutions to obtain high-resolution projections of ESMs include downscaling approaches that consider the information at a coarse-scale to make predictions at local scales. Complex and non-linear interdependence among local climate variables (e.g., temperature and precipitation) and large-scale predictors (e.g., pressure fields) motivate the use of neural network-based super-resolution architectures. In this work, we present auxiliary variables informed spatio-temporal neural architecture for statistical downscaling. The current study performs daily downscaling of precipitation variable from an ESM output at 1.15 degrees (~115 km) to 0.25 degrees (25 km) over the world's most climatically diversified country, India. We showcase significant improvement gain against three popular state-of-the-art baselines with a better ability to predict extreme events. To facilitate reproducible research, we make available all the codes, processed datasets, and trained models in the public domain.