Zobrazeno 1 - 9
of 9
pro vyhledávání: '"Maheshwary, Rishabh"'
Autor:
Dumitru, Razvan-Gabriel, Yadav, Vikas, Maheshwary, Rishabh, Clotan, Paul-Ioan, Madhusudhan, Sathwik Tejaswi, Surdeanu, Mihai
We present a simple variable quantization approach that quantizes different layers of a large language model (LLM) at different bit levels. Specifically, we quantize the most important layers to higher bit precision and less important layers to lower
Externí odkaz:
http://arxiv.org/abs/2406.17415
Autor:
Maheshwary, Rishabh, Yadav, Vikas, Nguyen, Hoang, Mahajan, Khyati, Madhusudhan, Sathwik Tejaswi
Instruction finetuning (IFT) is critical for aligning Large Language Models (LLMs) to follow instructions. While many effective IFT datasets have been introduced recently, they predominantly focus on high-resource languages like English. To better al
Externí odkaz:
http://arxiv.org/abs/2406.16783
Autor:
Pattnaik, Pulkit, Maheshwary, Rishabh, Ogueji, Kelechi, Yadav, Vikas, Madhusudhan, Sathwik Tejaswi
Direct Preference Optimization (DPO) is an effective technique that leverages pairwise preference data (usually one chosen and rejected response pair per user prompt) to align LLMs to human preferences. In practice, multiple responses can exist for a
Externí odkaz:
http://arxiv.org/abs/2403.07230
Autor:
Dancette, Corentin, Whitehead, Spencer, Maheshwary, Rishabh, Vedantam, Ramakrishna, Scherer, Stefan, Chen, Xinlei, Cord, Matthieu, Rohrbach, Marcus
Despite advances in Visual Question Answering (VQA), the ability of models to assess their own correctness remains underexplored. Recent work has shown that VQA models, out-of-the-box, can have difficulties abstaining from answering when they are wro
Externí odkaz:
http://arxiv.org/abs/2306.08751
Existing Math Word Problem (MWP) solvers have achieved high accuracy on benchmark datasets. However, prior works have shown that such solvers do not generalize well and rely on superficial cues to achieve high performance. In this paper, we first con
Externí odkaz:
http://arxiv.org/abs/2205.00177
Standard accuracy metrics have shown that Math Word Problem (MWP) solvers have achieved high performance on benchmark datasets. However, the extent to which existing MWP solvers truly understand language and its relation with numbers is still unclear
Externí odkaz:
http://arxiv.org/abs/2109.05925
Existing black box search methods have achieved high success rate in generating adversarial attacks against NLP models. However, such search methods are inefficient as they do not consider the amount of queries required to generate adversarial attack
Externí odkaz:
http://arxiv.org/abs/2109.04775
We study an important and challenging task of attacking natural language processing models in a hard label black box setting. We propose a decision-based attack strategy that crafts high quality adversarial examples on text classification and entailm
Externí odkaz:
http://arxiv.org/abs/2012.14956
We study an important task of attacking natural language processing models in a black box setting. We propose an attack strategy that crafts semantically similar adversarial examples on text classification and entailment tasks. Our proposed attack fi
Externí odkaz:
http://arxiv.org/abs/2012.13339