MiA-Signature: Approximating Global Activation for Long-Context Understanding
Yuqing Li Jiangnan Li Mo Yu Zheng Lin, Weiping Wang Jie Zhou · 2026
Read MiA-Signature: Approximating Global Activation for Long-Context Understanding by Yuqing Li Jiangnan Li Mo Yu Zheng Lin, Weiping Wang Jie Zhou (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-06416.txt.
Description
Li and five colleagues compress the brain-like activation behind a query into a usable signal. Cognitive science links reportable conscious access to global ignition across spread-out memory systems, yet people cannot directly access or list all that is activated. The authors propose the Mindscape Activation Signature, a compact signal that mimics the global influence of activation on later processing. In language-model systems it is built by submodular selection of high-level concepts covering the activated context. Light iterative updates from working memory can refine it.
Used as a conditioning signal, the signature mimics what the full activation state does while staying cheap to run. Adding it to retrieval-augmented and agentic systems yields consistent gains across long-context understanding tasks. The work turns a cognitive intuition into an engineering trick: carry the gist of global activation, not the whole activated store.
Catalog record
| CREATOR | Yuqing Li Jiangnan Li Mo Yu Zheng Lin, Weiping Wang Jie Zhou |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.06416