REVIEW 24 cited by
Landmark Attention: Random-Access Infinite Context Length for Transformers
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Landmark Attention: Random-Access Infinite Context Length for Transformers
read the original abstract
While Transformers have shown remarkable success in natural language processing, their attention mechanism's large memory requirements have limited their ability to handle longer contexts. Prior approaches, such as recurrent memory or retrieval-based augmentation, have either compromised the random-access flexibility of attention (i.e., the capability to select any token in the entire context) or relied on separate mechanisms for relevant context retrieval, which may not be compatible with the model's attention. In this paper, we present a novel approach that allows access to the complete context while retaining random-access flexibility, closely resembling running attention on the entire context. Our method uses a landmark token to represent each block of the input and trains the attention to use it for selecting relevant blocks, enabling retrieval of blocks directly through the attention mechanism instead of by relying on a separate mechanism. Our approach seamlessly integrates with specialized data structures and the system's memory hierarchy, enabling processing of arbitrarily long context lengths. We demonstrate that our method can obtain comparable performance with Transformer-XL while significantly reducing the number of retrieved tokens in each step. Finally, we show that fine-tuning LLaMA 7B with our method successfully extends its context length capacity to over 32k tokens, allowing for inference at the context lengths of GPT-4. We release the implementation of landmark attention and the code to reproduce our experiments at https://github.com/epfml/landmark-attention/.
Forward citations
Cited by 24 Pith papers
-
LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding
LongBench is the first bilingual multi-task benchmark for long context understanding in LLMs, containing 21 datasets in 6 categories with average lengths of 6711 words (English) and 13386 characters (Chinese).
-
LOCKS: Page-Local Compact Key Summaries for Efficient Long-Context Decoding
Page-local rank-8 spectral key summaries let sparse decode selection track the exact mass oracle and match FullKV quality at ~2% attended tokens with 2× latency cut at 1M context.
-
Anti-Periodic Positional Encoding: M\"obius Boundary Conditions Make In-Context Retrieval Reliable
An anti-periodic rotary positional encoding (M\u00f6bius RoPE) on 25% of heads collapses seed-to-seed variance in needle-in-a-haystack retrieval at matched perplexity.
-
Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression
KV cache compression causes task-dependent degradation in high-density reasoning due to disrupted CoT links; ShotKV mitigates this by preserving few-shot examples as indivisible semantic units through phase separation...
-
LongRoPE: Extending LLM Context Window Beyond 2 Million Tokens
LongRoPE extends LLM context windows to 2048k tokens via search for non-uniform positional interpolation, progressive fine-tuning from 256k, and short-context readjustment.
-
Extending Context Window of Large Language Models via Positional Interpolation
Position Interpolation linearly down-scales position indices to extend RoPE context windows to 32768 tokens with 1000-step fine-tuning, delivering strong long-context results on LLaMA 7B-65B while preserving short-con...
-
From Tensor Buffer to Distributed Memory Hierarchy: A Survey of KV Cache Management for LLM Serving
KV-cache serving systems concentrate into five archetypes under a four-axis taxonomy, with ownership explaining residual distributed design variance and seven measurement gaps blocking next steps.
-
Understanding Is Done Early: A Depth Division of Labor in Large Language Models and Its Use for Unbounded-Context Memory
Long-context memory can be organized by caching intermediate residual states and recomputing only query-conditioned upper layers over a bounded retrieved pack.
-
Raven: High-Recall Sequence Modeling with Sparse Memory Routing
Raven is a linear-time sequence model whose sparse, input-dependent routing writes tokens into dedicated memory slots, preserving long-context recall and extrapolating 16x beyond training length.
-
Remembering Distinct Items, Not Tokens: A Learnable Dirichlet-Process Cache Between State-Space Models and Attention
A DP-means allocate-on-novelty cache matches full-attention associative recall while storing only distinct items, and a minimal novelty gate recovers the rule end-to-end.
-
What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents
KV-cache eviction, prompt compression, recurrent state bounding, and agent memory consolidation are unified as one rate-distortion problem with a shared lower bound, shared failure mode, and transferable mechanisms.
-
A Hippocampus for Linear Attention: An Exact Memory for What the Recurrent State Forgets
HOLA pairs a compressive delta-rule recurrent state with a residual-selected exact KV cache and decoupled RMSNorm-gamma read, yielding lower perplexity than both standard linear attention and full-attention baselines ...
-
HPP: Hierarchical Programmatic Probing for Long Video Understanding by Decoupling Perception and Reasoning
HPP decouples perception from reasoning in long-video VLMs by having an LLM run iterative programmatic probes on hierarchically segmented video, reporting gains on LongVideoBench, EgoSchema, VideoMME, and MLVU.
-
CoDe-R: Refining Decompiler Output with LLMs via Rationale Guidance and Adaptive Inference
Rationale-guided fine-tuning plus dual-path adaptive inference lifts a 1.3B decompiler refiner to 50% average re-executability, a new lightweight SOTA on HumanEval-Decompile.
-
CoDe-R: Refining Decompiler Output with LLMs via Rationale Guidance and Adaptive Inference
CoDe-R refines LLM decompiler output via rationale-guided semantic injection and dynamic fallback inference, making a 1.3B model the first to exceed 50% average re-executability on HumanEval-Decompile.
-
IceCache: Memory-efficient KV-cache Management for Long-Sequence LLMs
IceCache combines semantic token clustering with PagedAttention to keep only 25% of the KV cache tokens while retaining 99% accuracy on LongBench and matching or beating prior offloading methods in latency.
-
Gated KalmaNet: A Fading Memory Layer Through Test-Time Ridge Regression
Gated KalmaNet uses exact Kalman gain computation with adaptive gating and Chebyshev iteration to improve SSM performance on long-context tasks over prior approximations like DeltaNet.
-
KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
KIVI applies asymmetric 2-bit quantization to KV cache with per-channel keys and per-token values, reducing memory 2.6x and boosting throughput up to 3.47x with near-identical quality on Llama, Falcon, and Mistral.
-
YaRN: Efficient Context Window Extension of Large Language Models
YaRN extends the context window of RoPE-based LLMs like LLaMA more efficiently than prior methods, using 10x fewer tokens and 2.5x fewer steps while surpassing state-of-the-art performance and enabling extrapolation b...
-
Latent Recurrent Transformer: Architecture Exploration, Training Strategies, and Scaling Behavior
Latent Recurrent Transformer augments autoregressive transformers with a cross-layer recurrent latent pathway from prior hidden states and uses interleaved parallel training to improve loss and in-context learning at ...
-
Episodic-Semantic Memory Architecture for Long-Horizon Scientific Agents
A dual-process memory architecture for scientific AI agents maintains 70-85% accuracy over 15,000 messages by using a constant 10-message episodic window and domain-specific semantic consolidation, consuming 62% fewer...
-
Mixtral of Experts
Mixtral 8x7B is a sparse MoE LLM activating 2 of 8 experts per layer that matches or exceeds Llama 2 70B and GPT-3.5 on benchmarks while using only 13B active parameters.
-
SAGE-32B: Agentic Reasoning via Iterative Distillation
SAGE-32B improves multi-tool agentic success rates over same-size baselines by combining iterative distillation with an inverse-reasoning meta-cognition head.
-
The Rise and Potential of Large Language Model Based Agents: A Survey
The paper surveys the origins, frameworks, applications, and open challenges of AI agents built on large language models.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.