[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85564-en":3,"doc-seo-85564-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},85564,16904993612988,"Olivia Brown","https://ap-avatar.wpscdn.com/davatar_a8503ba1806abce46bf441b54a3ca4cd",8,"Research & Report","From Words to Widgets for Controllable LLM Generation","Natural language remains the predominant way people interact with large language models (LLMs), yet subjective preferences like tone, style, and emphasis are hard to express precisely through prompting. Malleable Prompting converts preference expressions in prompts into interactive GUI widgets—sliders, dropdowns, and toggles—so users can configure values directly. A decoding algorithm steers token probabilities accordingly, and the interface visualizes each control’s influence for attribution and iteration comparison. A user study shows improved precision, controllability, and transparency over natural-language prompting.","From Words to Widgets for Controllable LLM Generation  \nChao Zhang∗ [cz468@cornell.edu](cz468@cornell.edu)[ ](cz468@cornell.edu)Cornell University Ithaca, NY, USA  \nYiren Liu∗ [yirenl2@illinois.edu](yirenl2@illinois.edu)[ ](yirenl2@illinois.edu)University of Illinois Urbana-Champaign Champaign, IL, USA  \nLunyiu Nie∗ [lynie@utexas.edu](lynie@utexas.edu)[ ](lynie@utexas.edu)The University of Texas at Austin Austin, TX, USA  \nJeffrey M. Rzeszotarski  \n[jrzeszotarski@loyola.edu](jrzeszotarski@loyola.edu)[ ](jrzeszotarski@loyola.edu)Loyola University Maryland Baltimore, MD, USA  \nYun Huang  \n[yunhuang@illinois.edu](yunhuang@illinois.edu)[ ](yunhuang@illinois.edu)University of Illinois Urbana-Champaign Champaign, IL, USA  \nTal August [taugust@illinois.edu](taugust@illinois.edu)[ ](taugust@illinois.edu)University of Illinois Urbana-Champaign Champaign, IL, USA  \narXiv :2604 . 10925v2 [ cs .HC] 10 Jul 2026  \nFigure 1: Malleable Prompting. Instead of repeatedly refining outputs through underspecified natural-language prompts (left), Malleable Prompting lets users reify prompt preferences as GUI widgets and directly manipulate them to steer generation (A) . Hovering over a widget reveals the output spans influenced by that preference, making individual effects inspectable (B) .  \nAbstract  \nNatural language remains the predominant way people interact with large language models (LLMs) . However, users often struggle to precisely express and control subjective preferences (e.g., tone, style, and emphasis) through prompting. We propose Malleable Prompting, a new interactive prompting technique for controllable LLM generation. It reifies preference expressions in natural language prompts into GUI widgets (e.g., sliders, dropdowns, and toggles) that users can directly configure to steer generation, while visualizing each control’s influence on the output to support attribution and comparison across iterations. To enable this interaction, we introduce an LLM decoding algorithm that modulates the token  \n∗ These authors contributed equally to this work.  \nThis work is licensed under a Creative Commons Attribution 4 .0 International License.  \nprobability distribution during generation based on preference expressions and their widget values. Through a user study, we show that Malleable Prompting helps participants achieve target preferences more precisely and is perceived as more controllable and transparent than natural language prompting alone.  \nCCS Concepts  \n• Human-centered computing → Natural language interfaces;  \n• Computing methodologies → Natural language processing.  \nKeywords  \nPrompting, Large Language Models, Human-AI Interaction  \n1 Introduction  \nWhen interacting with large language models (LLMs), users provide instructions via prompts, yet models rarely generate content  \nthat aligns perfectly with user intent on the first try. As a result, users often engage in an iterative process [35], refining the model’s output through turn-by-turn conversations to better match their preferences. For example, a user may ask a model to draft a blog post comparing two phones, then iterate by specifying comparison aspects (e.g., price, battery, aesthetics), altering the structure (bulleted vs. narrative), and fine-tuning style (e.g.,“more concise,”“less formal”) . While prior work has helped users revise concrete aspects of generated text, such as replacing words with synonyms, through direct manipulation [33], users still struggle to articulate and control more subjective preferences—such as tone, style, and emphasis—through prompting alone [26, 46] . This challenge stems from two interaction gulfs in natural language (NL) prompting:  \n• First, there is a gulf of execution [18]: users may not have clear preferences in mind at the outset, and articulating nuanced preferences in language is notoriously difficult [51] . Even when users do have a preference, it is still difficult for them to align model output with their expectations for the de","cbCairCxgeCb5KBB","https://ap.wps.com/l/cbCairCxgeCb5KBB","pdf",5460068,1,19,"English","en",105,"# Abstract\n# Introduction\n## Interaction gulfs in NL prompting\n## GUI widgets for preference control\n## Contributions","[{\"question\":\"What problem does Malleable Prompting address in natural-language prompting?\",\"answer\":\"Users struggle to express and control subjective preferences precisely, and they also find evaluation difficult as chat histories grow. Malleable Prompting targets both the execution gulf and the evaluation gulf.\"},{\"question\":\"How does Malleable Prompting make preferences controllable?\",\"answer\":\"It reifies preference expressions from natural-language prompts into GUI widgets such as sliders, dropdowns, and toggles, letting users directly set preference values to steer generation.\"},{\"question\":\"How does the system help users understand the effect of each preference?\",\"answer\":\"An interface visualizes each widget’s influence on the generated output, enabling attribution and comparison across iterative changes. A decoding algorithm modulates token probabilities based on preference expressions and widget values.\"}]",1784204622,48,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"from-words-to-widgets-for-controllable-llm-generation","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/from-words-to-widgets-for-controllable-llm-generation/85564/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-24","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does Malleable Prompting address in natural-language prompting?","Question",{"text":75,"@type":76},"Users struggle to express and control subjective preferences precisely, and they also find evaluation difficult as chat histories grow. Malleable Prompting targets both the execution gulf and the evaluation gulf.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does Malleable Prompting make preferences controllable?",{"text":80,"@type":76},"It reifies preference expressions from natural-language prompts into GUI widgets such as sliders, dropdowns, and toggles, letting users directly set preference values to steer generation.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the system help users understand the effect of each preference?",{"text":84,"@type":76},"An interface visualizes each widget’s influence on the generated output, enabling attribution and comparison across iterative changes. A decoding algorithm modulates token probabilities based on preference expressions and widget values.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":21,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":106,"slug":137},"General","general"]