6 papers
Exposing Long-Tail Safety Failures in Large Language Models through Efficient Diverse Response Sampling
Suvadeep Hajra, Palash Nandi, Tanmoy Chakraborty
Safety tuning through supervised fine-tuning and reinforcement learning from human feedback has substantially improved the robustness of large language models. However, it typicall…
MMA-ASIA: A Multilingual and Multimodal Alignment Framework for Culturally-Grounded Evaluation
Weihua Zheng, Zhengyuan Liu, Tanmoy Chakraborty +32
Large language models (LLMs) are now used worldwide, yet their multimodal understanding and reasoning often degrade outside Western, high-resource settings. We propose MMA-ASIA, a…
SABER: Uncovering Vulnerabilities in Safety Alignment via Cross-Layer Residual Connection
Maithili Joshi, Palash Nandi, Tanmoy Chakraborty
Large Language Models (LLMs) with safe-alignment training are powerful instruments with robust language comprehension capabilities. These models typically undergo meticulous alignm…
Innocence in the Crossfire: Roles of Skip Connections in Jailbreaking Visual Language Models
Palash Nandi, Maithili Joshi, Tanmoy Chakraborty
Language models are highly sensitive to prompt formulations - small changes in input can drastically alter their output. This raises a critical question: To what extent can prompt…
SAFE-MEME: Structured Reasoning Framework for Robust Hate Speech Detection in Memes
Palash Nandi, Shivam Sharma, Tanmoy Chakraborty
Memes act as cryptic tools for sharing sensitive ideas, often requiring contextual knowledge to interpret them correctly. It makes multimodal meme moderation difficult, as existing…
Recent Advances in Hate Speech Moderation: Multimodality and the Role of Large Models
Ming Shan Hee, Shivam Sharma, Rui Cao +4
In the evolving landscape of online communication, moderating hate speech (HS) presents an intricate challenge, compounded by the multimodal nature of digital content. This compreh…