16 citations · 35 across the 33 of their papers we have counts for
31 papers · 1 filter
DirectSpeech2LLM: A Simple End-to-End Framework to Mitigate Prompt Overfitting in Speech-LLMs
Hemant Yadav, Sunayana Sitaram, Roger Zimmermann +1
Speech-LLMs often exhibit prompt overfitting, where models solely trained on automatic speech recognition (ASR) instruction fail to generalize to new instructions such as speech tr…
Actions Speak Louder than Words: Measuring Cross-Lingual Policy Retention in Tool-Using Agents
Sourabrata Mukherjee, Kalika Bali, Sunayana Sitaram
When a tool-using agent is given the same task in a different language, does it still take the same steps? Multilingual evaluation rarely asks: it compares final answers and discar…
Pluralis v0.1: Towards a Multicultural, Multimodal, Multilingual Benchmark for AI Risk and Reliability
Alicia Parrish, Rajat Shinde, Sanket Badhe +57
Current AI safety evaluation and benchmarking frameworks predominantly rely on Western-centric culture-agnostic defaults that mask critical regional laws, socio-linguistic nuances,…
The Geometry of LLM-as-Judge: Why Inter-LLM Consensus Is Not Human Alignment
Sourabrata Mukherjee, Hamna Hamna, Kalika Bali +1
LLM judges now score most open-ended NLP output, and their mutual agreement is routinely read as evidence that the scores can be trusted. That reading is unsafe: judges may agree b…
DEPART: DEcomposing PARiTy across Multilingual LLMs
Manan Uppadhyay, Prashant Kodali, Pranjal Chitale +3
Multilingual Large Language Models (mLLMs) leaderboards report per-language accuracy but rarely explain why disparities emerge, leaving systemic biases unattributed and offering pr…
Building Benchmarks from the Ground Up: Community-Centered Evaluation of LLMs in Healthcare Chatbot Settings
Hamna Hamna, Gayatri Bhat, Sourabrata Mukherjee +5
Large Language Models (LLMs) are typically evaluated through general or domain-specific benchmarks testing capabilities that often lack grounding in the lived realities of end user…
UPDESH: Synthesizing Grounded Instruction Tuning Data for 13 Indic Languages
Pranjal A. Chitale, Varun Gumma, Sanchit Ahuja +4
Developing culturally grounded multilingual AI systems remains challenging, particularly for low-resource languages. While synthetic data offers promise, its effectiveness in multi…
How Deep Is Representational Bias in LLMs? The Cases of Caste and Religion
Agrima Seth, Monojit Choudhary, Sunayana Sitaram +3
Representational bias in large language models (LLMs) has predominantly been measured through single-response interactions and has focused on Global North-centric identities like r…
Fluent but Foreign: Even Regional LLMs Lack Cultural Alignment
Dhruv Agarwal, Anya Shukla, Sunayana Sitaram +1
Large language models (LLMs) are used worldwide, yet exhibit Western cultural tendencies. Many countries are now building ``regional'' or ``sovereign'' LLMs, but it remains unclear…
Uncovering inequalities in new knowledge learning by large language models across different languages
Chenglong Wang, Haoyu Tang, Xiyuan Yang +8
As large language models (LLMs) gradually become integral tools for problem solving in daily life worldwide, understanding linguistic inequality is becoming increasingly important.…
A Multilingual, Culture-First Approach to Addressing Misgendering in LLM Applications
Sunayana Sitaram, Adrian de Wynter, Isobel McCrum +2
Misgendering is the act of referring to someone by a gender that does not match their chosen identity. It marginalizes and undermines a person's sense of self, causing significant…
Contamination Report for Multilingual Benchmarks
Sanchit Ahuja, Varun Gumma, Sunayana Sitaram
Benchmark contamination refers to the presence of test datasets in Large Language Model (LLM) pre-training or post-training data. Contamination can lead to inflated scores on bench…
Exploring Pretraining via Active Forgetting for Improving Cross Lingual Transfer for Decoder Language Models
Divyanshu Aggarwal, Ashutosh Sathe, Sunayana Sitaram
Large Language Models (LLMs) demonstrate exceptional capabilities in a multitude of NLP tasks. However, the efficacy of such models to languages other than English is often limited…
Exploring Continual Fine-Tuning for Enhancing Language Ability in Large Language Model
Divyanshu Aggarwal, Sankarshan Damle, Navin Goyal +2
A common challenge towards the adaptability of Large Language Models (LLMs) is their ability to learn new languages over time without hampering the model's performance on languages…
HEALTH-PARIKSHA: Assessing RAG Models for Health Chatbots in Real-World Multilingual Settings
Varun Gumma, Ananditha Raghunath, Mohit Jain +1
Assessing the capabilities and limitations of large language models (LLMs) has garnered significant interest, yet the evaluation of multiple models in real-world scenarios remains…
JOOCI: a Framework for Learning Comprehensive Speech Representations
Hemant Yadav, Rajiv Ratn Shah, Sunayana Sitaram
Information in speech can be categorized into two groups: Content (what is being said, such as linguistics) and Other (how it is expressed such as information about speaker and par…
M5 -- A Diverse Benchmark to Assess the Performance of Large Multimodal Models Across Multilingual and Multicultural Vision-Language Tasks
Florian Schneider, Sunayana Sitaram
Since the release of ChatGPT, the field of Natural Language Processing has experienced rapid advancements, particularly in Large Language Models (LLMs) and their multimodal counter…
Speech Representation Learning Revisited: The Necessity of Separate Learnable Parameters and Robust Data Augmentation
Hemant Yadav, Sunayana Sitaram, Rajiv Ratn Shah
Speech modeling methods learn one embedding for a fixed segment of speech, typically in between 10-25 ms. The information present in speech can be divided into two categories: "wha…
Improving Self Consistency in LLMs through Probabilistic Tokenization
Ashutosh Sathe, Divyanshu Aggarwal, Sunayana Sitaram
Prior research has demonstrated noticeable performance gains through the use of probabilistic tokenizations, an approach that involves employing multiple tokenizations of the same…
sPhinX: Sample Efficient Multilingual Instruction Fine-Tuning Through N-shot Guided Prompting
Sanchit Ahuja, Kumar Tanmay, Hardik Hansrajbhai Chauhan +9
Despite the remarkable success of large language models (LLMs) in English, a significant performance gap remains in non-English languages. To address this, we introduce a novel app…
Cultural Conditioning or Placebo? On the Effectiveness of Socio-Demographic Prompting
Sagnik Mukherjee, Muhammad Farid Adilazuarda, Sunayana Sitaram +3
Socio-demographic prompting is a commonly employed approach to study cultural biases in LLMs as well as for aligning models to certain cultures. In this paper, we systematically pr…
Beyond Metrics: Evaluating LLMs' Effectiveness in Culturally Nuanced, Low-Resource Real-World Scenarios
Millicent Ochieng, Varun Gumma, Sunayana Sitaram +5
The deployment of Large Language Models (LLMs) in real-world applications presents both opportunities and challenges, particularly in multilingual and code-mixed communication sett…
Teaching LLMs to Abstain across Languages via Multilingual Feedback
Shangbin Feng, Weijia Shi, Yike Wang +6
Multilingual LLMs often have knowledge disparities across languages, with larger gaps in under-resourced languages. Teaching LLMs to abstain in the face of knowledge gaps is thus a…
PARIKSHA: A Large-Scale Investigation of Human-LLM Evaluator Agreement on Multilingual and Multi-Cultural Data
Ishaan Watts, Varun Gumma, Aditya Yadavalli +3
Evaluation of multilingual Large Language Models (LLMs) is challenging due to a variety of factors -- the lack of benchmarks with sufficient linguistic diversity, contamination of…