Paper 2605.06416
MiA-Signature: Approximating Global Activation for Long-Context Understanding
- Published
- May 2026
- Research lab
- Independent
- Citations
- 1
- GitHub
- Not linked
01 In brief
Summary
The paper introduces Mindscape Activation Signature (MiA-Signature), a compact representation of the global activation pattern induced by a query over a semantic memory space, inspired by cognitive science theories of global ignition and partial access.
It is constructed via submodular selection of high-level concepts (session summaries) that cover the activated context, optionally refined iteratively.
The signature serves as a conditioning signal for retrieval and generation, complementing local evidence.
The method is instantiated in two settings: static RAG (one-shot signature) and an iterative agent (evolving signature).
Experiments on long-context benchmarks (DetectiveQA, NarrativeQA, NovelHopQA, NoCha) show consistent improvements over query-only baselines.
Key results: conditioning retrieval on the signature improves average R@10 by 10.9% and task performance by 3.8% in static RAG; in the agent setting, the signature improves retrieval recall on all benchmarks.
The signature is more beneficial for retrieval than for generation; its answer-time value is selective, helping when global constraints are needed.
Coverage-aware submodular initialization outperforms First-K in static RAG.
Query rewriting is benchmark-dependent, kept fixed on NovelHopQA.
02 From the paper
Abstract
A growing body of work in cognitive science suggests that reportable conscious access is associated with \emph{global ignition} over distributed memory systems, while such activation is only partially accessible as individuals cannot directly access or enumerate all activated contents. This tension suggests a plausible mechanism that cognition may rely on a compact representation that approximates the global influence of activation on downstream processing. Inspired by this idea, we introduce the concept of \textbf{Mindscape Activation Signature (MiA-Signature)}, a compressed representation of the global activation pattern induced by a query. In LLM systems, this is instantiated via submodular-based selection of high-level concepts that cover the activated context space, optionally refined through lightweight iterative updates using working memory. The resulting MiA-Signature serves as a conditioning signal that approximates the effect of the full activation state while remaining computationally tractable. Integrating MiA-Signatures into both RAG and agentic systems yields consistent performance gains across multiple long-context understanding tasks.