{"total":86,"rows":[{"id":"2606.19348","title":"DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence","created":"2026-04-26","cited_by_count":1040,"area":"moe"},{"id":"2604.08224","title":"Externalization in LLM Agents: A Unified Review of Memory, Skills, Protocols and Harness Engineering","created":"2026-04-09","cited_by_count":74,"area":"agents"},{"id":"2608.30320","title":"On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability","created":"2026-08-31","cited_by_count":35,"area":"moe"},{"id":"2603.09697","title":"Mousse: Rectifying the Geometry of Muon with Curvature-Aware Preconditioning","created":"2026-03-10","cited_by_count":17,"area":null},{"id":"2601.21204","title":"Scaling Embeddings Outperforms Scaling Experts in Language Models","created":"2026-01-29","cited_by_count":15,"area":"moe"},{"id":"2605.20613","title":"HRM-Text: Efficient Pretraining Beyond Scaling","created":"2026-05-20","cited_by_count":9,"area":null},{"id":"2602.10652","title":"UMEM: Unified Memory Extraction and Management Framework for Generalizable Memory","created":"2026-02-11","cited_by_count":8,"area":null},{"id":"2602.11683","title":"ThinkRouter: Efficient Reasoning via Routing Thinking between Latent and Discrete Spaces","created":"2026-02-12","cited_by_count":8,"area":"lr"},{"id":"2603.05484","title":"Towards Multimodal Lifelong Understanding: A Dataset and Agentic Baseline","created":"2026-03-05","cited_by_count":8,"area":"agents"},{"id":"2602.03359","title":"MeKi: Memory-based Expert Knowledge Injection for Efficient LLM Scaling","created":"2026-02-03","cited_by_count":7,"area":null},{"id":"2607.12385","title":"PM-Bench: Evaluating Prospective Memory in LLM Agents","created":"2026-07-14","cited_by_count":7,"area":"agents"},{"id":"2603.23516","title":"MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokens","created":"2026-03-06","cited_by_count":6,"area":null},{"id":"2603.28342","title":"Kernel-Smith: A Unified Recipe for Evolutionary Kernel Optimization","created":"2026-03-30","cited_by_count":6,"area":null},{"id":"2603.09983","title":"MoE-SpAc: Efficient MoE Inference Based on Speculative Activation Utility in Heterogeneous Edge Scenarios","created":"2026-02-12","cited_by_count":5,"area":"moe"},{"id":"2605.20948","title":"Memory Grafting: Scaling Language Model Pre-training via Offline Conditional Memory","created":"2026-05-20","cited_by_count":5,"area":"moe"},{"id":"2604.02309","title":"go-$m$HC: Direct Parameterization of Manifold-Constrained Hyper-Connections via Generalized Orthostochastic Matrices","created":"2026-04-02","cited_by_count":4,"area":null},{"id":"2605.00814","title":"Persistent Visual Memory: Sustaining Perception for Deep Generation in LVLMs","created":"2026-05-01","cited_by_count":4,"area":"mllm"},{"id":"2607.25380","title":"Memory for Large Language Models","created":"2026-07-28","cited_by_count":4,"area":null},{"id":"2602.04706","title":"LiteToken: Removing Intermediate Merge Residues From BPE Tokenizers","created":"2026-02-04","cited_by_count":3,"area":null},{"id":"2604.08519","title":"Cram Less to Fit More: Training Data Pruning Improves Memorization of Facts","created":"2026-04-09","cited_by_count":3,"area":null},{"id":"2605.21463","title":"Mem-$\\pi$: Adaptive Memory through Learning When and What to Generate","created":"2026-05-20","cited_by_count":3,"area":null},{"id":"2605.24869","title":"Lngram: N-gram Conditional Memory in Latent Space","created":"2026-05-24","cited_by_count":3,"area":null},{"id":"2606.08347","title":"Tensorizing Engram: Sharing Latents Across N-Gram Embeddings is Beneficial in LLMs","created":"2026-06-06","cited_by_count":3,"area":null},{"id":"2606.19172","title":"User as Engram: Internalizing Per-User Memory as Local Parametric Edits","created":"2026-06-17","cited_by_count":3,"area":"lr"},{"id":"2607.25614","title":"MemSFT: Mitigating Alignment Tax with an External Parametric Memory","created":"2026-07-28","cited_by_count":3,"area":"alignment"},{"id":"2608.17050","title":"Frozen Memory Is Not Enough: Rethinking External Memory as Extraction","created":"2026-08-17","cited_by_count":3,"area":null},{"id":"2601.16531","title":"A Collision-Free Hot-Tier Extension for Engram-Style Conditional Memory: A Controlled Study of Training Dynamics","created":"2026-01-23","cited_by_count":2,"area":null},{"id":"2601.21461","title":"L$^3$: Large Lookup Layers","created":"2026-01-29","cited_by_count":2,"area":"moe"},{"id":"2602.13680","title":"AllMem: A Memory-centric Recipe for Efficient Long-context Modeling","created":"2026-02-14","cited_by_count":2,"area":null},{"id":"2603.15202","title":"Simple is Better: Multiplication May Be All You Need for LLM Request Scheduling","created":"2026-03-16","cited_by_count":2,"area":"inference"},{"id":"2604.27263","title":"Decoupling the Benefits of Subword Tokenization for Language Model Training via Byte-level Simulation","created":"2026-04-29","cited_by_count":2,"area":null},{"id":"2605.06546","title":"Efficient Pre-Training with Token Superposition","created":"2026-05-07","cited_by_count":2,"area":null},{"id":"2605.06615","title":"When and Why SignSGD Outperforms SGD: A Theoretical Study Based on $\\ell_1$-norm Lower Bounds","created":"2026-05-07","cited_by_count":2,"area":null},{"id":"2605.25680","title":"Simulating Human Memory with Language Models","created":"2026-05-25","cited_by_count":2,"area":null},{"id":"2606.01155","title":"When Data Is Scarce: Scaling Sparse Language Models with Repeated Training","created":"2026-05-31","cited_by_count":2,"area":null},{"id":"2606.28116","title":"Mechanism-Driven Monitors for Preemptive Detection of LLM Training Instability","created":"2026-06-26","cited_by_count":2,"area":null},{"id":"2608.29327","title":"When to Adapt: Conditional Memory Adapters for Retention-Preserving Domain Specialization","created":"2026-08-29","cited_by_count":2,"area":null},{"id":"2601.22203","title":"Beyond Conditional Computation: Retrieval-Augmented Genomic Foundation Models with Gengram","created":"2026-01-29","cited_by_count":1,"area":"rag"},{"id":"2603.10006","title":"Adaptive Engram Memory System for Indonesian Language Model: Generative AI Based on TOBA LM for Batak and Minang Language","created":"2026-02-17","cited_by_count":1,"area":null},{"id":"2603.21096","title":"Mixture of Chapters: Scaling Learnt Memory in Transformers","created":"2026-03-22","cited_by_count":1,"area":null},{"id":"2603.25414","title":"Decidable By Construction: Design-Time Verification for Truly Fearless Systems","created":"2026-03-26","cited_by_count":1,"area":null},{"id":"2604.11297","title":"The Past Is Not Past: Memory-Enhanced Dynamic Reward Shaping","created":"2026-04-13","cited_by_count":1,"area":null},{"id":"2604.21724","title":"Beyond N-gram: Data-Aware X-GRAM Extraction for Efficient Embedding Parameter Scaling","created":"2026-04-23","cited_by_count":1,"area":null},{"id":"2605.10739","title":"Geospatial-Temporal Sensemaking of Remote Sensing Activity Detections with Multimodal Large Language Model","created":"2026-05-11","cited_by_count":1,"area":"mllm"},{"id":"2605.12426","title":"Geometric Factual Recall in Transformers","created":"2026-05-12","cited_by_count":1,"area":null},{"id":"2605.13179","title":"Does Engram Do Memory Retrieval in Autoregressive Image Generation?","created":"2026-05-13","cited_by_count":1,"area":null},{"id":"2605.13370","title":"Phasor Memory Networks: Stable Backpropagation Through Time for Scalable Explicit Memory","created":"2026-05-13","cited_by_count":1,"area":null},{"id":"2605.16893","title":"NGM: A Plug-and-Play Training-Free Memory Module for LLMs","created":"2026-05-16","cited_by_count":1,"area":null},{"id":"2605.20309","title":"Tiny-Engram: Trigger-Indexed Concept Tables for Generative Vision","created":"2026-05-19","cited_by_count":1,"area":"diffusion"},{"id":"2605.27686","title":"Tensor Memory: Fixed-Size Recurrent State for Long-Horizon Transformers","created":"2026-05-26","cited_by_count":1,"area":null},{"id":"2606.02780","title":"Do Value Vectors in Deep Layers Need Context from the Residual Stream?","created":"2026-06-01","cited_by_count":1,"area":null},{"id":"2606.12113","title":"Augmenting Molecular Language Models with Local $n$-gram Memory","created":"2026-06-10","cited_by_count":1,"area":null},{"id":"2607.04969","title":"Train Smarter, Not Longer: Memorization-Guided Data Reuse for Efficient LLM Training","created":"2026-07-06","cited_by_count":1,"area":null},{"id":"2607.20518","title":"CANN Bench: Benchmarking Agent Generated Kernels against Real NPU and Algorithmic Limits","created":"2026-07-08","cited_by_count":1,"area":null},{"id":"2608.05745","title":"UniVVT: A Unified End-to-End Framework for High-Fidelity Video Virtual Try-on","created":"2026-08-06","cited_by_count":1,"area":null},{"id":"2609.28399","title":"Memory Attention","created":"2026-09-23","cited_by_count":1,"area":null},{"id":"2601.10810","title":"Digital Metabolism: Decoupling Logic from Facts via Regenerative Unlearning -- Towards a Pure Neural Logic Core","created":"2026-01-15","cited_by_count":0,"area":"lr"},{"id":"2602.02499","title":"ROSA-Tuning: Enhancing Long-Context Modeling via Suffix Matching","created":"2026-01-14","cited_by_count":0,"area":null},{"id":"2604.06956","title":"NestPipe: Large-Scale Recommendation Training on 1,500+ Accelerators via Nested Pipelining","created":"2026-04-08","cited_by_count":0,"area":null},{"id":"2605.04400","title":"Contextual Memory-Enhanced Source Coding for Low-SNR Communications","created":"2026-05-06","cited_by_count":0,"area":null},{"id":"2605.13297","title":"PaMM: Periodic Motif Memory for Atomistic Models with an Explicit Local-Structure Interface","created":"2026-05-13","cited_by_count":0,"area":null},{"id":"2605.18556","title":"Key-Gram: Extensible World Knowledge for Embodied Manipulation","created":"2026-05-18","cited_by_count":0,"area":null},{"id":"2605.18848","title":"Exact Linear Attention","created":"2026-05-13","cited_by_count":0,"area":"moe"},{"id":"2606.09888","title":"SinkRec: Mitigating Semantic State Sink in Long Sequence Recommendation with Memory-Conditioned Gated Delta Networks","created":"2026-06-03","cited_by_count":0,"area":null},{"id":"2607.07707","title":"Co-LMLM: Continuous-Query Limited Memory Language Models","created":"2026-07-08","cited_by_count":0,"area":null},{"id":"2607.16254","title":"More Than Memory: Task-Conditioned Signed FFN Writes in Long-Context Retrieval","created":"2026-06-27","cited_by_count":0,"area":null},{"id":"2607.16305","title":"LookME: Lookup-Based Multimodal Embeddings for Layer Injection in Vision-Language Models","created":"2026-07-14","cited_by_count":0,"area":"mllm"},{"id":"2607.18413","title":"Convolution for Large Language Models","created":"2026-07-20","cited_by_count":0,"area":null},{"id":"2608.12325","title":"Position: Reasoning is a Learnable Rule-Based Process","created":"2026-05-29","cited_by_count":0,"area":null},{"id":"2608.12419","title":"LoKiFormer: Locality-aware Attention with Decoupled Knowledge Memory for Efficient Large Language Model Pretraining","created":"2026-08-12","cited_by_count":0,"area":"moe"},{"id":"2608.26112","title":"TreeGraft: Adaptive Multi-Drafter Grafting for Tree-Based Speculative Decoding","created":"2026-05-28","cited_by_count":0,"area":"inference"},{"id":"2609.03426","title":"Lngram v2: Latent N-Gram Memory with Interpretable Discrete Representations","created":"2026-09-03","cited_by_count":0,"area":null},{"id":"2609.13478","title":"BadEngram: Backdoor Attack on Gated Memory Components in LLMs","created":"2026-09-11","cited_by_count":0,"area":"safety"},{"id":"2609.15126","title":"MoME: Mixture-of-Memory Embeddings for Context-Aware Sparse Lookup","created":"2026-09-14","cited_by_count":0,"area":"moe"},{"id":"2609.25802","title":"Latest Exact Match Attention","created":"2026-09-22","cited_by_count":0,"area":null},{"id":"2609.25853","title":"MemoryAthena: Adaptive Routing over Latent and Generated Memories","created":"2026-09-22","cited_by_count":0,"area":null},{"id":"2609.33270","title":"Structured Sparse Memory for Recurrent Reasoning","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.34633","title":"GenMem: Generative Symbolic Memory for Self-Evolving Harness","created":"2026-09-28","cited_by_count":0,"area":"agents"},{"id":"2609.35578","title":"FactorEngram: Factorized N-gram Memory with Basis-Level Gating for Language Models","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35663","title":"Late Attention Layers Alone Can Copy Entity Tokens, but Not Without Attending to Their Context","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.36301","title":"MoRE: Scaling mixture of experts with hardware-aware low-rank routing","created":"2026-09-28","cited_by_count":0,"area":"moe"},{"id":"2609.37198","title":"V-Engram: Trigger-Indexed External Memory for Modular Text-to-Image Personalization","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2609.39661","title":"The Evolution of Attention in Large Language Models: Mechanisms, Trade-offs, and Emerging Trends","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2610.00394","title":"Forking: Sudden Overfitting Under Replay","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2610.12599","title":"STAPLE: Static Sparse Upcycling of Pretrained Checkpoints via Embedding Modules","created":"2026-10-08","cited_by_count":0,"area":"moe"},{"id":"2610.20456","title":"The COUNT: Counting N-grams Unlocks Scaling on Host Memory and Training-Free Adaptation","created":"2026-10-14","cited_by_count":0,"area":null}],"next_cursor":null}