Recent research demonstrates that pretrained large language models (LLMs) require minimal but carefully curated data for instruction-tuning. Building on LLMs, multi-modal models further contend with far larger data volumes, heterogeneous sources, and cross-modal alignment, amplifying both the necessity and the complexi...
Mengyao Lyu, Yan Li, Huasong Zhong et al.· 0 citations
Hate speech detectors must separate hateful posts from the far more common offensive but non-hateful ones, and two popular ideas for doing so are hierarchical (coarse-to-fine) classification and sentiment features. We revisit our earlier transformer-based framework, which combines both ideas, and test each component in...
Language Models (LMs) have demonstrated exceptional performance across various Natural Language Processing (NLP) tasks. Despite these advancements, LMs can inherit and amplify societal biases related to sensitive attributes such as gender and race, limiting their adoption in real-world applications. Therefore, fairness...
Zhipeng Yin, Zichong Wang, Avash Palikhe et al.· 0 citations
Scientific discovery often begins with an unexpected observation that prompts a new research question. In a world with altered physical laws, can large language models (LLMs) notice unexpected behavior, such as heavier objects falling more slowly, and discover the law through experiments without being told what to inve...
Zhen Yang, Wen Wang, Xander Xu et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Fine-tuning Large Language Models (LLMs) on instruction data can quietly erode their safety, even when no training example looks harmful. Because this safety-degrading data looks benign, standard content filters struggle to catch it. We tackle this problem by developing SimpleLoRIF, a scalable gradient embedding method...
S. Reddy, Amro Abbas, Anshuman Suri et al.· 0 citations
Communication between agents has mostly been studied on tasks with a shared objective and fast, clear feedback. Trading has neither: each agent manages its own account, with decisions being scored by markets late and noisily. A trading swarm is thus a population of independent decision-makers rather than a team, for wh...
Avid Fayaz, Harry Zhang, H{\aa}vard Alstadheim et al.· 0 citations
Visual world models increasingly support spatial reasoning by generating RGB views of unobserved states. However, RGB has become the default interface largely because it is well suited for visual generation, rather than necessarily being the most effective representation for downstream reasoning. Different spatial ques...
Zengqi Zhao, Zun Wang, Shoubin Yu et al.· 0 citations
This work proposes \method, an evaluation allocation framework that combines sparse search-time feedback with targeted final confirmation and achieves higher mean performance than full evaluation in most settings while using approximately half as many evaluation cells.
Si-Chun Luo, Yi Huang, Mu-Kai Li et al.· 0 citations
Safety alignment research for large language models (LLMs) remains concentrated on high-resource languages, leaving open how existing alignment methods perform in low-resource, culturally distinct settings. We study this question for Hausa, a language spoken by more than 200 million people across West Africa. We introd...
Valdini Douglace Lemofouet, Blessing Ngozi Uzor, P. Anyanwu et al.· 0 citations
Sparse mixture-of-experts (MoE) language-model training can exhibit large loss spikes with Muon-family optimizers. We introduce EnergyCal, a lightweight rule that attenuates routed-expert updates according to the normalized squared Frobenius norm of their finite-step Newton-Schulz directions. EnergyCal leaves the spect...
Direct Preference Optimization (DPO) typically relies on preference pairs generated from diverse model pools and ranked by an external LLM judge. The recently proposed delta-learning hypothesis suggests that comparable supervision can instead be obtained from a single pair of models by treating the larger model's respo...
G. Curci, S. Simonazzi, Andrea Molinari et al.· 0 citations
Multimodal perception forms the foundation of multimodal intelligence. In particular, spatial perception and multimodal long-context understanding are two pivotal tasks that underpin embodied AI and multimodal agents. However, synthesizing targeted, high-quality training data for these tasks remains difficult and costl...
Flint-interactive gives AI agents a language for creating interactive visualizations. Click a chart, filter data, highlight related content, and explore beyond the view the author originally chose. The post Flint-interactive: A language for creating interactive data visualizations appeared first on Microsoft Research.
