Megalodon: Efficient LLM Pretraining and Inference with Unlimited Context Length
Paper
•
2404.08801
•
Published
•
68
LM Architectures
-
-
RecurrentGemma: Moving Past Transformers for Efficient Open Language Models
Paper • 2404.07839 • Published • 48 -
Eagle and Finch: RWKV with Matrix-Valued States and Dynamic Recurrence
Paper • 2404.05892 • Published • 39 -
Mamba: Linear-Time Sequence Modeling with Selective State Spaces
Paper • 2312.00752 • Published • 143 -
Multi-Head Mixture-of-Experts
Paper • 2404.15045 • Published • 61 -
Jamba: A Hybrid Transformer-Mamba Language Model
Paper • 2403.19887 • Published • 111 -
KAN: Kolmogorov-Arnold Networks
Paper • 2404.19756 • Published • 113 -
Better & Faster Large Language Models via Multi-token Prediction
Paper • 2404.19737 • Published • 78 -
Contextual Position Encoding: Learning to Count What's Important
Paper • 2405.18719 • Published • 5 -
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality
Paper • 2405.21060 • Published • 68 -
An Image is Worth More Than 16x16 Patches: Exploring Transformers on Individual Pixels
Paper • 2406.09415 • Published • 52 -
Alleviating Distortion in Image Generation via Multi-Resolution Diffusion Models
Paper • 2406.09416 • Published • 30 -
Transformers meet Neural Algorithmic Reasoners
Paper • 2406.09308 • Published • 45 -
Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling
Paper • 2406.07522 • Published • 41 -
Explore the Limits of Omni-modal Pretraining at Scale
Paper • 2406.09412 • Published • 11 -
Accessing GPT-4 level Mathematical Olympiad Solutions via Monte Carlo Tree Self-refine with LLaMa-3 8B
Paper • 2406.07394 • Published • 29 -
VideoLLM-online: Online Video Large Language Model for Streaming Video
Paper • 2406.11816 • Published • 25 -
Mixture of A Million Experts
Paper • 2407.04153 • Published • 5 -
Scaling Retrieval-Based Language Models with a Trillion-Token Datastore
Paper • 2407.12854 • Published • 32 -
MoMa: Efficient Early-Fusion Pre-training with Mixture of Modality-Aware Experts
Paper • 2407.21770 • Published • 23 -
Transformer Explainer: Interactive Learning of Text-Generative Models
Paper • 2408.04619 • Published • 161 -
Jamba-1.5: Hybrid Transformer-Mamba Models at Scale
Paper • 2408.12570 • Published • 34 -
Show-o: One Single Transformer to Unify Multimodal Understanding and Generation
Paper • 2408.12528 • Published • 52 -
LLMs + Persona-Plug = Personalized LLMs
Paper • 2409.11901 • Published • 34 -
MonoFormer: One Transformer for Both Diffusion and Autoregression
Paper • 2409.16280 • Published • 18 -
Differential Transformer
Paper • 2410.05258 • Published • 180 -
Byte Latent Transformer: Patches Scale Better Than Tokens
Paper • 2412.09871 • Published • 106 -
OmniFlow: Any-to-Any Generation with Multi-Modal Rectified Flows
Paper • 2412.01169 • Published • 13 -
Monet: Mixture of Monosemantic Experts for Transformers
Paper • 2412.04139 • Published • 13 -
MH-MoE:Multi-Head Mixture-of-Experts
Paper • 2411.16205 • Published • 29 -
Hymba: A Hybrid-head Architecture for Small Language Models
Paper • 2411.13676 • Published • 46 -
SageAttention2 Technical Report: Accurate 4 Bit Attention for Plug-and-play Inference Acceleration
Paper • 2411.10958 • Published • 56 -
BitNet a4.8: 4-bit Activations for 1-bit LLMs
Paper • 2411.04965 • Published • 69 -
Mixture-of-Transformers: A Sparse and Scalable Architecture for Multi-Modal Foundation Models
Paper • 2411.04996 • Published • 52 -
rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking
Paper • 2501.04519 • Published • 280 -
MiniMax-01: Scaling Foundation Models with Lightning Attention
Paper • 2501.08313 • Published • 295 -
Tensor Product Attention Is All You Need
Paper • 2501.06425 • Published • 89 -
Transformer^2: Self-adaptive LLMs
Paper • 2501.06252 • Published • 55 -
Learnings from Scaling Visual Tokenizers for Reconstruction and Generation
Paper • 2501.09755 • Published • 37 -
FAST: Efficient Action Tokenization for Vision-Language-Action Models
Paper • 2501.09747 • Published • 25 -
Test-Time Preference Optimization: On-the-Fly Alignment via Iterative Textual Feedback
Paper • 2501.12895 • Published • 62 -
Autonomy-of-Experts Models
Paper • 2501.13074 • Published • 45 -
Parameters vs FLOPs: Scaling Laws for Optimal Sparsity for Mixture-of-Experts Language Models
Paper • 2501.12370 • Published • 11 -
Over-Tokenized Transformer: Vocabulary is Generally Worth Scaling
Paper • 2501.16975 • Published • 32 -
PixelWorld: Towards Perceiving Everything as Pixels
Paper • 2501.19339 • Published • 17 -
Scaling Embedding Layers in Language Models
Paper • 2502.01637 • Published • 24 -
FastKV: KV Cache Compression for Fast Long-Context Processing with Token-Selective Propagation
Paper • 2502.01068 • Published • 17 -
Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach
Paper • 2502.05171 • Published • 142 -
CMoE: Fast Carving of Mixture-of-Experts for Efficient LLM Inference
Paper • 2502.04416 • Published • 12 -
The Curse of Depth in Large Language Models
Paper • 2502.05795 • Published • 40 -
LM2: Large Memory Models
Paper • 2502.06049 • Published • 30 -
TransMLA: Multi-head Latent Attention Is All You Need
Paper • 2502.07864 • Published • 56 -
LLM Pretraining with Continuous Concepts
Paper • 2502.08524 • Published • 29 -
Large Language Diffusion Models
Paper • 2502.09992 • Published • 120 -
Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention
Paper • 2502.11089 • Published • 160 -
Continuous Diffusion Model for Language Modeling
Paper • 2502.11564 • Published • 54 -
Multimodal Mamba: Decoder-only Multimodal State Space Model via Quadratic to Linear Distillation
Paper • 2502.13145 • Published • 38 -
MoBA: Mixture of Block Attention for Long-Context LLMs
Paper • 2502.13189 • Published • 17 -
Token-Efficient Long Video Understanding for Multimodal LLMs
Paper • 2503.04130 • Published • 95 -
Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs
Paper • 2503.01743 • Published • 88 -
Transformers without Normalization
Paper • 2503.10622 • Published • 167 -
Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models
Paper • 2503.09573 • Published • 72 -
Forgetting Transformer: Softmax Attention with a Forget Gate
Paper • 2503.02130 • Published • 32 -
OmniMamba: Efficient and Unified Multimodal Understanding and Generation via State Space Models
Paper • 2503.08686 • Published • 19 -
RWKV-7 "Goose" with Expressive Dynamic State Evolution
Paper • 2503.14456 • Published • 152