{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2602-03784","title":"Fix the Structural Bottleneck: Context Compression via Explicit Information Transmission","arxiv_id":"2602.03784","date":"2026-02-03","proceeding":null,"authors":["Jiangnan Ye","Hanqi Yan","Zhenyi Shen","Heng Chang","Ye Mao","Yulan He"],"abstract":"Long-context LLM agents often struggle with growing token, memory, and latency costs, making efficient context compression essential for practical deployment. Existing LLM-as-a-compressor methods remain noticeably inferior to using the full context. We find that this gap partly stems from their inability to preserve contextual information effectively. In this work, we revisit context compression from a structural perspective and identify two key bottlenecks in standard LLM-based compressors: limited coordination among compression tokens during information aggregation, and layerwise dilution that weakens useful signals from intermediate hidden states. To address these limitations, we propose ComprExIT, a new context compression framework based on explicit information transmission. ComprExIT adaptively selects features across frozen LLM layers, then allocates information from anchors to compression slots through a globally coordinated transport plan. Experiments on 12 datasets show that ComprExIT consistently outperforms strong soft-compression baselines, improving average F1 by up to 18.5%, while adding only ~1% trainable parameters and achieving more than 2x faster compression than the fastest baselines. The code will be released upon acceptance.","url_abs":"https://arxiv.org/abs/2602.03784","url_pdf":"https://arxiv.org/pdf/2602.03784","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2602.03784","atlas_url":"https://app.syntology.ai/?focus=2602.03784","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2602.03784"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/Jiangnan0522/ComprExIT","reach":null}],"summary":{"unverified":5},"by_repo_kind":{"found_in_text":{"samples":5,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"6cfd975f730cf408","entry":"beacon_predict","repo":"Jiangnan0522/ComprExIT","repo_kind":"found_in_text","path":"src/model/inference.py","file_url":"https://github.com/Jiangnan0522/ComprExIT/blob/HEAD/src/model/inference.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6cfd975f730cf408"}},{"code_sha256_prefix":"9f88a25274d07db2","entry":"data_loading_factory","repo":"Jiangnan0522/ComprExIT","repo_kind":"found_in_text","path":"src/data_processing/data_loading.py","file_url":"https://github.com/Jiangnan0522/ComprExIT/blob/HEAD/src/data_processing/data_loading.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"9f88a25274d07db2"}},{"code_sha256_prefix":"1d56830e39983466","entry":"get_data_collator_factory","repo":"Jiangnan0522/ComprExIT","repo_kind":"found_in_text","path":"src/data_processing/ntp_preprocessing.py","file_url":"https://github.com/Jiangnan0522/ComprExIT/blob/HEAD/src/data_processing/ntp_preprocessing.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"1d56830e39983466"}},{"code_sha256_prefix":"392995da7f414f44","entry":"get_device_stats","repo":"Jiangnan0522/ComprExIT","repo_kind":"found_in_text","path":"src/device_utils.py","file_url":"https://github.com/Jiangnan0522/ComprExIT/blob/HEAD/src/device_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"392995da7f414f44"}},{"code_sha256_prefix":"414973f82aff6b26","entry":"supports_flash_attention_2","repo":"Jiangnan0522/ComprExIT","repo_kind":"found_in_text","path":"src/device_utils.py","file_url":"https://github.com/Jiangnan0522/ComprExIT/blob/HEAD/src/device_utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"414973f82aff6b26"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}