← Back to all news
arXiv AI

The Intrusion Pathway: How Non-Evidential Context Degrades Language Model Performance

The paper identifies that language models lose accuracy when prompts include non‑evidential context—irrelevant statements or random tokens that do not overlap semantically with the query. This degradation is traced to an "Intrusion Pathway" involving specific attention heads that amplify distractor content and weaken correct answers. By targeting and masking a small set of these intrusion heads, the study recovers most of the lost performance with minimal impact on useful context.

Related stories

arXiv AI
Sep 3

Language Models Can Control Their Own Attention

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
llmsefficiency
More like this →
arXiv Machine Learning
Jun 5

Masks Can Be Distracting: On Context Comprehension in Diffusion Language Models

arXiv:2511. 21338v2 Announce Type: replace Abstract: Masked Diffusion Language Models (MDLMs) have recently emerged as a promising alternative to Autoregressive Language Models (ARLMs), leveraging a denoising objective that, in principle, should enable more uniform context utilisation.

By Julianna Piskorz, Cristina Pinneri, Alvaro Correia, Motasem Alfarra, Risheek Garrepalli, Christos Louizos
llmsdiffusionfine-tuningsafety
More like this →
arXiv AI
Aug 25

Lexical Perturbations Disrupt LLM Reasoning: An Empirical Study of Attention Diversion

The study investigates how lexical perturbations—such as keyboard noise, character swaps, and filler insertion—affect large language models (LLMs) on reasoning benchmarks. Four open-weight instruction-tuned models and frontier models were evaluated, revealing that character-level perturbations significantly reduce accuracy, especially on multi-step reasoning tasks, while filler insertion has minimal impact. The authors attribute this asymmetry to Attention Diversion, where fragmented subword tokenization draws disproportionate attention in middle and final transformer layers; they demonstrate that both token content and attention allocation are coupled, making it difficult for inference-time repair strategies to fully recover performance.

By Jiaqian Zhu, Yang Zhang, Junhua Ding, Xiaowei Yu
llmsnlpbenchmarks
More like this →
arXiv AI
Sep 25

Near-Oracle KV Selection via Pre-hoc Sparsity for Long-Context Inference

The paper introduces Pre-hoc Sparsity (PrHS), a method that selects key-value (KV) cache entries before attention scoring to avoid posterior bias in large language model inference. By bounding mutual‑information loss through the dropped attention mass, PrHS offers explicit accuracy control and implements three orthogonal selectors across time, depth, and layer. Experiments on LLaMA and Mistral models show that PrHS cuts retrieval overhead by over 90%, achieves higher sparsity than HShare, and delivers significant speedups and reduced FLOPs on NVIDIA A100 GPUs while maintaining near‑dense accuracy.

By Yifei Gao, Lei Wang, Rong-Cheng Tu, Qixin Zhang, Jun Cheng, Dacheng Tao
llmsefficiencysafety
More like this →