# Google 与 KAIST 论文提出 Declarative Attention，让模型自选读取上下文，注意力开销降低 52.0%

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-09-05 11:13
- AIHOT 分数：37
- AIHOT 链接：https://aihot.virxact.com/items/cmtnu4sx40ca2roqs05mnl9t8
- 原文链接：https://x.com/rohanpaul_ai/status/2096074134287786223

## AI 摘要

Google 与 KAIST 的论文提出 Declarative Attention（DA），模型通过 <global>、<focus>、<local> 三种模式在思维链中声明需要关注的上下文，推理引擎据此跳过大部分 KV cache 读取，无需额外的外置打分器。

## 正文

New Google Deployment Paper shows instead of forcing an LLM to reread its entire context for every token, let the model choose what to read

cutting attention work by 52.0%

Today’s models often reread the entire stored context for every new token, even when only a small section matters.

Declarative Attention changes that: the model itself says which part of the context it needs, and the inference engine skips the rest.

No separate scorer is needed to search the whole context first.

Across 15 long-context tasks, attention work fell by 52.0% on Gemma-4-31B and 31.1% on Qwen-3.6-27B, while accuracy dropped by 1.27 and 2.75 percentage points.

Larger models handled this trade-off better, suggesting the approach could improve further with training.