paper-with-me

홈 › Papers

HELM: Hyperbolic Large Language Models via Mixture-of-Curvature Experts

2025-05-30 · Neil He, Rishabh Anand, Hiren Madhu, Ali Maatouk, Smita Krishnaswamy, Leandros Tassiulas, Menglin Yang, Rex Ying

Large language models (LLMs) have shown great success in text modeling tasks across domains. However, natural language exhibits inherent semantic hierarchies and nuanced geometric structure, which current LLMs do not capture completely owing to their reliance on Euclidean operations. Recent studies have also shown that not respecting the geometry of token embeddings leads to training instabilities and degradation of generative capabilities. These findings suggest that shifting to non-Euclidean geometries can better align language models with the underlying geometry of text. We thus propose to operate fully in Hyperbolic space, known for its expansive, scale-free, and low-distortion properties. We thus introduce HELM, a family of HypErbolic Large Language Models, offering a geometric rethinking of the Transformer-based LLM that addresses the representational inflexibility, missing set of necessary operations, and poor scalability of existing hyperbolic LMs. We additionally introduce a Mixture-of-Curvature Experts model, HELM-MICE, where each expert operates in a distinct curvature space to encode more fine-grained geometric structure from text, as well as a dense model, HELM-D. For HELM-MICE, we further develop hyperbolic Multi-Head Latent Attention (HMLA) for efficient, reduced-KV-cache training and inference. For both models, we develop essential hyperbolic equivalents of rotary positional encodings and RMS normalization. We are the first to train fully hyperbolic LLMs at billion-parameter scale, and evaluate them on well-known benchmarks such as MMLU and ARC, spanning STEM problem-solving, general knowledge, and commonsense reasoning. Our results show consistent gains from our HELM architectures -- up to 4% -- over popular Euclidean architectures used in LLaMA and DeepSeek, highlighting the efficacy and enhanced reasoning afforded by hyperbolic geometry in large-scale LM pretraining.

📄 PDF Abstract BibTeX arXiv:2505.24722

Code (1)

graph-and-geometric-learning/helm 공식 구현 pytorch

Tasks

ARCGeneral KnowledgeMMLU

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음
LLaMA LLaMA is a collection of foundation language models ranging from 7B to 65B parameters. It is based on the transformer architecture with various improvements that were…
ALIGN In the ALIGN method, visual and language representations are jointly trained from noisy image alt-text data. The image and text encoders are learned via contrastive loss…
SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

EEG-Based Multimodal Learning via Hyperbolic Mixture-of-Curvature Experts

2026-04-14 · Runhe Zhou, Shanglin Li, Guanxiang Huang, Xinliang Zhou 외 arxiv

Electroencephalography (EEG)-based multimodal learning integrates brain signals with complementary modalities to improve mental state assessment, providing great clinical potential. The effectiveness of such paradigms la…

Representation LearningEmotion Recognition

HyperHELM: Hyperbolic Hierarchy Encoding for mRNA Language Modeling

2025-09-29 · Max van Spengler, Artem Moskalev, Tommaso Mansi, Mangal Prakash 외 arxiv

Language models are increasingly applied to biological sequences like proteins and mRNA, yet their default Euclidean geometry may mismatch the hierarchical structures inherent to biological data. While hyperbolic geometr…

Adaptive Hyperbolic Kernels: Modulated Embedding in de Branges-Rovnyak Spaces

2025-11-13 · Leping Si, Meimei Yang, Hui Xue, Shipeng Zhu 외 arxiv

Hierarchical data pervades diverse machine learning applications, including natural language processing, computer vision, and social network analysis. Hyperbolic space, characterized by its negative curvature, has demons…

Parameter-Efficient Fine-Tuning of LLMs with Mixture of Space Experts

2026-02-16 · Buze Zhang, Jinkai Tao, Zilang Zeng, Neil He 외 arxiv

Large Language Models (LLMs) have achieved remarkable progress, with Parameter-Efficient Fine-Tuning (PEFT) emerging as a key technique for downstream task adaptation. However, existing PEFT methods mainly operate in Euc…

parameter-efficient fine-tuning

ACE-HGNN: Adaptive Curvature Exploration Hyperbolic Graph Neural Network

2021-10-15 · Xingcheng Fu, JianXin Li, Jia Wu, Qingyun Sun 외

Graph Neural Networks (GNNs) have been widely studied in various graph data mining tasks. Most existingGNNs embed graph data into Euclidean space and thus are less effective to capture the ubiquitous hierarchical structu…

Graph LearningGraph Neural NetworkMulti-agent Reinforcement LearningRepresentation Learning