长上下文论文趋势 2023–2026

长窗口建模

统计范围：NeurIPS、ICLR、ICML、CVPR 四个会议，2023–2026 年，共 838 篇被归入该主题的论文。

主要结论

每年论文数

各会议论文数

按年份和会议统计

长上下文论文数，按年份和会议
年份NeurIPSICLRICMLCVPR合计
20231818151566
202452436021176
202598875317255
2026130829831341

代表论文

按年份和会议均匀挑选，方便快速看到每一届的样貌

全部长上下文论文

年份全部2023202420252026
会议全部NeurIPSICLRICMLCVPR

共 255 篇，第 4 / 6 页

  1. ICML 2025Test-Time Training Provably Improves Transformers as In-context Learners
  2. ICML 2025TimeBridge: Non-Stationarity Matters for Long-term Time Series Forecasting
  3. ICML 2025Training Dynamics of In-Context Learning in Linear Attention
  4. ICML 2025TypyBench: Evaluating LLM Type Inference for Untyped Python Repositories
  5. ICML 2025Understanding and Improving Length Generalization in Recurrent Models
  6. ICML 2025Understanding Synthetic Context Extension via Retrieval Heads
  7. ICML 2025VideoRoPE: What Makes for Good Video Rotary Position Embedding?
  8. NeurIPS 2025A Controllable Examination for Long-Context Language Models
  9. NeurIPS 2025A Little Depth Goes a Long Way: The Expressive Power of Log-Depth Transformers
  10. NeurIPS 2025Absence Bench: Language Models Can’t See What’s Missing
  11. NeurIPS 2025AdmTree: Compressing Lengthy Context with Adaptive Semantic Trees
  12. NeurIPS 2025AttentionPredictor: Temporal Patterns Matter for KV Cache Compression
  13. NeurIPS 2025Block-Biased Mamba for Long-Range Sequence Processing
  14. NeurIPS 2025ChunkKV: Semantic-Preserving KV Cache Compression for Efficient Long-Context LLM Inference
  15. NeurIPS 2025Compress, Gather, and Recompute: REFORMing Long-Context Processing in Transformers
  16. NeurIPS 2025Constant Bit-size Transformers Are Turing Complete
  17. NeurIPS 2025Context-Aware Regularization with Markovian Integration for Attention-Based Nucleotide Analysis
  18. NeurIPS 2025ConTextTab: A Semantics-Aware Tabular In-Context Learner
  19. NeurIPS 2025CORAL: Disentangling Latent Representations in Long-Tailed Diffusion
  20. NeurIPS 2025DeepHalo: A Neural Choice Model with Controllable Context Effects
  21. NeurIPS 2025Disentangling Latent Shifts of In-Context Learning with Weak Supervision
  22. NeurIPS 2025DyG-Mamba: Continuous State Space Modeling on Dynamic Graphs
  23. NeurIPS 2025Eagle 2.5: Boosting Long-Context Post-Training for Frontier Vision-Language Models
  24. NeurIPS 2025Effects of Dropout on Performance in Long-range Graph Learning Tasks
  25. NeurIPS 2025Efficient Low Rank Attention for Long-Context Inference in Large Language Models
  26. NeurIPS 2025Efficient Prompt Compression with Evaluator Heads for Long-Context Transformer Inference
  27. NeurIPS 2025Enhancing the Maximum Effective Window for Long-Term Time Series Forecasting
  28. NeurIPS 2025FLAME: Fast Long-context Adaptive Memory for Event-based Vision
  29. NeurIPS 2025FlashMD: long-stride, universal prediction of molecular dynamics
  30. NeurIPS 2025FlexSelect: Flexible Token Selection for Efficient Long Video Understanding
  31. NeurIPS 2025From Softmax to Score: Transformers Can Effectively Implement In-Context Denoising Steps
  32. NeurIPS 2025Hardware-aligned Hierarchical Sparse Attention for Efficient Long-term Memory Access
  33. NeurIPS 2025Hierachical Balance Packing: Towards Efficient Supervised Fine-tuning for Long-Context LLM
  34. NeurIPS 2025Hierarchical Demonstration Order Optimization for Many-shot In-Context Learning
  35. NeurIPS 2025Hierarchical Implicit Neural Emulators
  36. NeurIPS 2025HiFC: High-efficiency Flash-based KV Cache Swapping for Scaling LLM Inference
  37. NeurIPS 2025Homogeneous Keys, Heterogeneous Values: Exploiting Local KV Cache Asymmetry for Long-Context LLMs
  38. NeurIPS 2025HoPE: Hybrid of Position Embedding for Long Context Vision-Language Models
  39. NeurIPS 2025How Data Mixing Shapes In-Context Learning: Asymptotic Equivalence for Transformers with MLPs
  40. NeurIPS 2025In-Context Compositional Learning vis Sparse Coding Transformer
  41. NeurIPS 2025In-Context Learning Strategies Emerge Rationally
  42. NeurIPS 2025KeyDiff: Key Similarity-Based KV Cache Eviction for Long-Context LLM Inference in Resource-Constrained Environments
  43. NeurIPS 2025Know Thyself by Knowing Others: Learning Neuron Identity from Population Context
  44. NeurIPS 2025KVzip: Query-Agnostic KV Cache Compression with Context Reconstruction
  45. NeurIPS 2025L-MTP: Leap Multi-Token Prediction Beyond Adjacent Context for Large Language Models
  46. NeurIPS 2025L$^2$M: Mutual Information Scaling Law for Long-Context Language Modeling
  47. NeurIPS 2025Language Models (Mostly) Know When to Stop Reading
  48. NeurIPS 2025Learning to Focus: Causal Attention Distillation via Gradient‐Guided Token Pruning
  49. NeurIPS 2025Let Me Think! A Long Chain of Thought Can Be Worth Exponentially Many Short Ones
  50. NeurIPS 2025Linear Transformers Implicitly Discover Unified Numerical Algorithms

在 Claude、Cursor 里直接问这个趋势

接上 Paperlist MCP 后，写综述、挑方向或准备投稿时，不用离开对话框就能拿到这页的数字：每年各会议的论文数、增长倍数、涨得最快的子方向和代表论文，AI 会按你的问题整理。

可以直接复制这句话问 AI

用 paperlist 查长上下文在 NeurIPS 2023–2026 的论文趋势，哪一年涨得最快？

AI 会调用的工具：get_topic_stats {"topic": "long-context"}

两分钟接入 Paperlist MCP（Claude Code、Cursor、Codex、ChatGPT）

常见问题

2023–2026 年四大会议一共有多少篇长上下文论文？
共 838 篇，按年份分别是 2023 年 66 篇、2024 年 176 篇、2025 年 255 篇、2026 年 341 篇。
NeurIPS 2026 有多少篇长上下文论文？
NeurIPS 2026 有 130 篇长上下文论文，上一届是 98 篇。
ICLR 2026 有多少篇长上下文论文？
ICLR 2026 有 82 篇长上下文论文，上一届是 87 篇。
ICML 2026 有多少篇长上下文论文？
ICML 2026 有 98 篇长上下文论文，上一届是 53 篇。
CVPR 2026 有多少篇长上下文论文？
CVPR 2026 有 31 篇长上下文论文，上一届是 17 篇。
哪个会议的长上下文论文最多？
NeurIPS，2023–2026 年共 298 篇。

「长上下文与记忆」下的其他主题