Imagen de Portada

When Less is More: The LLM Scaling Paradox in Context Compression

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Guo, Ruishan, Liu, Yibing, Ma, Guoxin, Wang, Yan, Zhang, Yueyang, Xia, Long, Chen, Kecheng, Sun, Zhiyuan, Shi, Daiting
Formato:	Preprint
Publicado:	2026
Materias:	Machine Learning
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

Thinking as Compression: Your Reasoning Model is Secretly a Context Compressor
por: Ma, Guoxin, et al.
Publicado: (2026)

ConsistRM: Improving Generative Reward Models via Consistency-Aware Self-Training
por: Liang, Yu, et al.
Publicado: (2026)

Advancing General-Purpose Reasoning Models with Modular Gradient Surgery
por: Cai, Min, et al.
Publicado: (2026)

TRE: Encouraging Exploration in the Trust Region
por: Huang, Chao, et al.
Publicado: (2026)

Moirai 2.0: When Less Is More for Time Series Forecasting
por: Liu, Chenghao, et al.
Publicado: (2025)

LIMR: Less is More for RL Scaling
por: Li, Xuefeng, et al.
Publicado: (2025)

Less Is More, but Where? Dynamic Token Compression via LLM-Guided Keyframe Prior
por: Li, Yulin, et al.
Publicado: (2025)

Less-to-More Generalization: Unlocking More Controllability by In-Context Generation
por: Wu, Shaojin, et al.
Publicado: (2025)

Rethinking Tokenization for Clinical Time Series: When Less is More
por: Attrach, Rafi Al, et al.
Publicado: (2025)

Less or More From Teacher: Exploiting Trilateral Geometry For Knowledge Distillation
por: Hu, Chengming, et al.
Publicado: (2023)

ReflectRM: Boosting Generative Reward Models via Self-Reflection within a Unified Judgment Framework
por: Qin, Kai, et al.
Publicado: (2026)

Less Is More: Elevating RAG via Performance-Driven Context Compression
por: Cui, Ziqiang, et al.
Publicado: (2025)

When More is Less: Understanding Chain-of-Thought Length in LLMs
por: Wu, Yuyang, et al.
Publicado: (2025)

When Less Latent Leads to Better Relay: Information-Preserving Compression for Latent Multi-Agent LLM Collaboration
por: Li, Yiping, et al.
Publicado: (2026)

Less is More: Benchmarking LLM Based Recommendation Agents
por: Chauhan, Kargi, et al.
Publicado: (2026)

Quantize What Counts: More for Keys, Less for Values
por: Hariri, Mohsen, et al.
Publicado: (2025)

Less Is More? When Dataset Context Hurts LLM-Generated Dataset Descriptions
por: Gan, Lisa-Yao, et al.
Publicado: (2026)

Less is More: Empowering GUI Agent with Context-Aware Simplification
por: Chen, Gongwei, et al.
Publicado: (2025)

When Less Is More: Binary Feedback Can Outperform Ordinal Comparisons in Ranking Recovery
por: Xu, Shirong, et al.
Publicado: (2025)

Cut Less, Fold More: Model Compression through the Lens of Projection Geometry
por: Saukh, Olga, et al.
Publicado: (2026)

Input-Time Scaling: Adding Noise and Irrelevance into Less-Is-More Drastically Improves Reasoning Performance and Efficiency
por: Huang, Rapheal, et al.
Publicado: (2025)

Learn More, Forget Less: A Gradient-Aware Data Selection Approach for LLM
por: Liu, Yibai, et al.
Publicado: (2025)

When Less is More: On the Value of "Co-training" for Semi-Supervised Software Defect Predictors
por: Majumder, Suvodeep, et al.
Publicado: (2022)

Less is More: Towards Simple Graph Contrastive Learning
por: Zhao, Yanan, et al.
Publicado: (2025)

When Less is More: Achieving Faster Convergence in Distributed Edge Machine Learning
por: Basani, Advik Raj, et al.
Publicado: (2024)

Less Noise, More Voice: Reinforcement Learning for Reasoning via Instruction Purification
por: Guo, Yiju, et al.
Publicado: (2026)

Two-stage LLM Fine-tuning with Less Specialization and More Generalization
por: Wang, Yihan, et al.
Publicado: (2022)

Less is More: on the Over-Globalizing Problem in Graph Transformers
por: Xing, Yujie, et al.
Publicado: (2024)

When Do "More Contexts" Help with Sarcasm Recognition?
por: Nimase, Ojas, et al.
Publicado: (2024)

When Less Is More.
por: Lettis, Lucy
Publicado: (1998)

Code Less, Align More: Efficient LLM Fine-tuning for Code Generation with Data Pruning
por: Tsai, Yun-Da, et al.
Publicado: (2024)

When Less is More: 8-bit Quantization Improves Continual Learning in Large Language Models
por: Zhang, Michael S., et al.
Publicado: (2025)

Language Models, Graph Searching, and Supervision Adulteration: When More Supervision is Less and How to Make More More
por: Frydenlund, Arvid
Publicado: (2025)

Less is More: Improving LLM Alignment via Preference Data Selection
por: Deng, Xun, et al.
Publicado: (2025)

APB: Accelerating Distributed Long-Context Inference by Passing Compressed Context Blocks across GPUs
por: Huang, Yuxiang, et al.
Publicado: (2025)

Less is More: Optimizing Function Calling for LLM Execution on Edge Devices
por: Paramanayakam, Varatheepan, et al.
Publicado: (2024)

Less Approximates More: Harmonizing Performance and Confidence Faithfulness via Hybrid Post-Training for High-Stakes Tasks
por: Ma, Haokai, et al.
Publicado: (2026)

Less is More: Accurate Speech Recognition & Translation without Web-Scale Data
por: Puvvada, Krishna C., et al.
Publicado: (2024)

Expand More, Shrink Less: Shaping Effective-Rank Dynamics for Dense Scaling in Recommendation
por: Li, Guoming, et al.
Publicado: (2026)

Learning More with Less: A Dynamic Dual-Level Down-Sampling Framework for Efficient Policy Optimization
por: Wang, Chao, et al.
Publicado: (2025)