The paper introduces the concept of stale binding, where large language models (LLMs) answer with outdated values despite having newer information in context. It presents Controlled In-Context Memory (CICM), a benchmark to track and test the use of updated information in conversations and agent logs. Experiments on open‑source models reveal an attention drift mechanism that favors old values, and the authors propose a simple attention‑redirecting intervention that largely corrects these errors without harming correct answers.
By Junyu Guo, Yuchen Fang, Shangding Gu, Costas Spanos, James Demmel, Javad Lavaei
As preferences, goals, and facts change, LLM agents must use the current state while earlier versions remain in context. Yet they can answer with an old value of the same variable, a failure that we c...
arXiv:2607. 01002v1 Announce Type: cross Abstract: In long-context use, large language models frequently synthesize answers from the meaning of a relevant context span rather than literally copy-pasting them.
By Aryo Pradipta Gema, Beatrice Alex, Pasquale Minervini
The paper introduces Declarative Attention (DA), a protocol that lets language models explicitly declare which parts of their context to focus on during generation. By partitioning decoding into full-context, region-specific, and recent-output-only modes, the inference engine can skip large portions of the KV cache, dramatically reducing attended tokens. Experiments on 15 long-context tasks with off-the-shelf models show significant savings (52.0% and 31.1% reductions) with only modest accuracy drops that diminish as model size increases.
By Namgyu Ho, Huzama Ahmad, Woosung Koh, Se-Young Yun, Tal Schuster, Cicero Nogueira dos Santos
Hybrid language models combine attention with a fixed-size recurrent state, yet the distinct roles of each component are not well understood. The authors introduce two cache-level interventions—split-prefill and state-swap—to isolate the contributions of the KV cache (attention) and the recurrent state. Experiments on Qwen3.5 and Falcon-H1 show that exact retrieval depends almost entirely on attention, while output language and persona rely mainly on recurrence, with the state-swap intervention confirming that answers derive from the KV side and language from the recurrent side.
By Kirill Afendulev, Alexey Dontsov, Elena Tutubalina, Anton Korznikov
The paper investigates why Mamba models underperform Transformers on in-context retrieval tasks. It challenges the common belief that Mamba’s fixed-size recurrent state limits retrieval, showing that even when the query precedes the context, Mamba still lags behind Transformers. The study reveals that Mamba’s selection gates can strongly influence output, but they rely on relevance signals that only emerge gradually with depth, causing a retrieval bottleneck beyond state capacity.
By Nir Yarden, Amit Ben-Artzy, Roy Schwartz
arXiv:2512. 14391v3 Announce Type: replace-cross Abstract: In-context learning is fundamental to modern Large Language Models (LLMs); however, prevailing architectures impose a rigid and fixed contextual structure by assigning linear or constant positional indices.
By Huayang Li, Tianyu Zhao, Deng Cai, Richard Sproat
arXiv:2601.12075v2 Announce Type: replace
Abstract: Language models used in retrieval-augmented settings must arbitrate between parametric knowledge stored in their weights and contextual information...
By Mehrdad Farahani, Franziska Penzkofer, Richard Johansson
The paper investigates how the dimension of attention heads influences long‑context retrieval in language models. Experiments show that while pretraining loss and short‑context tasks are largely unaffected, wider heads significantly improve retrieval performance at fixed model sizes. Ablation studies indicate that this benefit persists even when attention weights are frozen, suggesting that wider heads enhance adaptation of other model components such as embeddings and MLPs.
By Seiya Tokui, Shoichiro Yamaguchi, Sosuke Kobayashi
The paper investigates where and how large language models encode signals that indicate answer correctness. By examining hidden states, token probabilities, residual-stream features, attention, and their combinations, the authors find that correctness signals are concentrated in the answer span and that different signal families complement each other. Fusing these signals improves robustness, especially under distribution shifts, and can be used to control retrieval in downstream tasks.
By Yuansen Liu, Yixuan Tang, Anthony Kum Hoe Tung
arXiv:2609.35663v2 Announce Type: replace
Abstract: Large language models (LLMs) reliably perform entity copying, in which a model copies tokens referring to an entity, termed entity tokens, from the...
By Muyu He, Yuchen Liu, Ran Tao, Li Zhang
arXiv:2606. 10650v1 Announce Type: cross Abstract: The scalability of Large Language Models (LLMs) to long contexts is fundamentally constrained by the quadratic complexity of standard attention, motivating the adoption of linear attention mechanisms with sub-quadratic cost.
By Xin Wang, Hui Shen, Boyuan Zheng, Xueshen Liu, Minkyoung Cho, Zhongwei Wan, Zesen Zhao, Zhuoqing Mao, Shen Yan, Mi Zhang