Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.04496 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915983690563584 |
|---|---|
| author | Zhang, Zhenliang Wang, Wenqing Hu, Yong Yang, Yaming Gao, Jiaheng Shen, Chen Wan, Xiaojun |
| author_facet | Zhang, Zhenliang Wang, Wenqing Hu, Yong Yang, Yaming Gao, Jiaheng Shen, Chen Wan, Xiaojun |
| contents | Long-Text Understanding (LTU) at million-token scale requires balancing reasoning fidelity with computational efficiency. Frontier long-context LLMs can process millions of token contexts end-to-end, but they suffer from high token consumption and attention dilution. In parallel, specialized LTU agents often sacrifice fidelity through task-agnostic abstractions like graph construction or indexing. We identify a key insight for LTU: query-relevant information is typically sparse relative to the full document, so effective reasoning should rely on a query-sufficient subset rather than the entire context. To address this, we propose SCOUT, a new paradigm for LTU that shifts from passive processing to active information foraging. It treats the document as an explorable environment and answers from a compact, provenance-grounded epistemic state. Guided by state-level gap diagnosis, SCOUT adaptively alternates between coarse-to-fine exploration and anchored state updates that progressively contract its epistemic state toward query sufficiency. Experiments show that SCOUT matches state-of-the-art proprietary models while reducing token consumption by up to 8x. Moreover, SCOUT remains stable as context length scales, substantially alleviating the practical cost-performance trade-off. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_04496 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | SCOUT: Active Information Foraging for Long-Text Understanding with Decoupled Epistemic States Zhang, Zhenliang Wang, Wenqing Hu, Yong Yang, Yaming Gao, Jiaheng Shen, Chen Wan, Xiaojun Computation and Language Long-Text Understanding (LTU) at million-token scale requires balancing reasoning fidelity with computational efficiency. Frontier long-context LLMs can process millions of token contexts end-to-end, but they suffer from high token consumption and attention dilution. In parallel, specialized LTU agents often sacrifice fidelity through task-agnostic abstractions like graph construction or indexing. We identify a key insight for LTU: query-relevant information is typically sparse relative to the full document, so effective reasoning should rely on a query-sufficient subset rather than the entire context. To address this, we propose SCOUT, a new paradigm for LTU that shifts from passive processing to active information foraging. It treats the document as an explorable environment and answers from a compact, provenance-grounded epistemic state. Guided by state-level gap diagnosis, SCOUT adaptively alternates between coarse-to-fine exploration and anchored state updates that progressively contract its epistemic state toward query sufficiency. Experiments show that SCOUT matches state-of-the-art proprietary models while reducing token consumption by up to 8x. Moreover, SCOUT remains stable as context length scales, substantially alleviating the practical cost-performance trade-off. |
| title | SCOUT: Active Information Foraging for Long-Text Understanding with Decoupled Epistemic States |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2605.04496 |