Hierarchical Instruction Conditioning for Controlled Long-Document Summarization

Hierarchical Instruction Conditioning for Controlled Long-Document Summarization | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 2 March 2026 V1 Latest version Share on Hierarchical Instruction Conditioning for Controlled Long-Document Summarization Authors : Haowen Shi 0009-0003-0362-9031 [email protected] and Yichen Zong Authors Info & Affiliations https://doi.org/10.22541/au.177247853.38733832/v1 135 views 89 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Despite significant advancements in Large Language Models (LLMs) for long-document summarization, current models often lack precise control when faced with complex, multi-faceted user instructions, manifesting as content focus misalignment and constraint violation. To address these critical limitations, we propose Hierarchical Instruction Conditioning for Summarization (HICS), a novel framework designed to enhance the fine-grained control capabilities of LLMs. HICS systematically decomposes complex user instructions into "Hierarchical Instruction Primitives" (categorized by core goal, focus areas, style, and constraints) and strategically injects these primitives incrementally at different stages of the LLM's decoding process. The framework integrates an Instruction Parser, Hierarchical Instruction Embedding, and Staged Primitive Injection to dynamically condition generation. Built upon an adapted LLaMA-2-70B backbone, HICS is trained and evaluated on two new datasets: DocSum-ControlSet and LongSum-Eval. Our extensive experiments demonstrate that HICS-Large significantly outperforms strong baselines, including LLaMA-2-70B with standard fine-tuning and GPT-3.5-Turbo, achieving superior scores on ROUGE-L, Focus Score, and Constraint Adherence. Ablation studies and human evaluations further validate its exceptional ability to generate high-quality summaries that precisely align with complex user instructions. Supplementary Material File (hics.pdf) Download 1.86 MB Information & Authors Information Version history V1 Version 1 02 March 2026 Copyright This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License Keywords complex instructions fine-grained control instruction conditioning large language models summarization Authors Affiliations Haowen Shi 0009-0003-0362-9031 [email protected] Sichuan University of Science and Engineering View all articles by this author Yichen Zong Sichuan University of Science and Engineering View all articles by this author Metrics & Citations Metrics Article Usage 135 views 89 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Haowen Shi, Yichen Zong. Hierarchical Instruction Conditioning for Controlled Long-Document Summarization. Authorea . 02 March 2026. DOI: https://doi.org/10.22541/au.177247853.38733832/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.177247853.38733832/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fe49f1a0f9edfa9',t:'MTc3OTIwOTY0NQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();