[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85315-en":3,"doc-seo-85315-105":29,"detail-sidebar-cat-0-en-105":90},{"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":4,"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},85315,1099514068365,"Aurelia","https://ap-avatar.wpscdn.com/avatar/10000253d8d9f28188e?_k=1776742907772140068",8,"Research & Report","PRISM Edit: One Vector for All Temporal Answers","Model editing updates large language models without retraining, but temporal facts expose a flaw in locate-and-edit: new information should be current while old information can remain historically correct. Using causal tracing, the work shows LLMs implement temporal distinction via two-stage internal computation: early MLP layers retrieve a time-agnostic subject representation, and later layers apply temporal modulation to produce time-correct answers. Based on this, PRISM Edit optimizes a single polysemous representation across time contexts without architectural changes.","PRISM Edit: One Vector for All Temporal Answers*  \nChen Huang1,†, Qi Zheng1,†, Ruiqin Zheng2 , Long Zeng1,* , Yuantong Xu2,*  \n1Tsinghua University, 2ByteDance  \n†Equal contribution. * Corresponding authors.  \nCorrespondence: [zenglong@sz.tsinghua.edu.cn](zenglong@sz.tsinghua.edu.cn)  \narXiv :2607 . 1 1327v 1 [ cs .LG] 13 Jul 2026  \nAbstract  \nModel editing keeps large language models (LLMs) up to date without retraining, but temporal facts expose a limitation of the prevailing locate-and-edit paradigm: an update is not always a replacement. When a fact changes, the new answer should become current while the old answer may remain correct in historical time contexts. Building on this insight, we use causal tracing to show that LLMs already support this distinction via a two-stage internal computation: early MLP layers retrieve a time-agnostic subject representation, and later layers modulate it with temporal context to yield the time-correct answer. Motivated by this finding, we introduce PRISM Edit, which optimizes a single polysemous representation across temporal contexts and leverages the model’s inherent modulation pathway to route it to temporally correct predictions—without any architectural modification. We evaluate on TIMECONFLICT, a new temporal editing benchmark we introduce, and on temporally augmented COUNTERFACT. PRISM Edit improves over the best baseline by +23.3 Temporal Consistency (TC) and +33.7 Current Relative-time Score (CRS) on average while being more than 2 × faster. Code and data are publicly available in an anonymized repository at [https://anonymous.4open.science/r/](https://anonymous.4open.science/r/)[ ](https://anonymous.4open.science/r/)PRISM-Edit-ABE2 .  \n1 Introduction  \nLarge language models store extensive factual knowledge, yet the world they describe keeps changing. Knowledge editing allows models to continuously acquire new knowledge without full retraining, typically by modifying a small set of parameters associated with a target memory (Cao et al., 2021 ; Mitchell et al., 2022 ; Meng et al., 2022, 2023 ; Zhang et al., 2024 ; Fang et al., 2025) . However, existing methods implicitly assume that every  \n* Code and data: [https://github.com/Cheer-Huang/](https://github.com/Cheer-Huang/)[ ](https://github.com/Cheer-Huang/)PRISM-Edit  \nedit is a wholesale replacement of the old fact. This assumption breaks down for temporal facts, where the old answer does not become wrong—it merely ceases to be current. For example, when the U.S. presidency transfers from Biden to Trump in January 2025, the new fact should become current, yet “Who was the U.S. president in 2023?” must still return Biden. A temporal editor must therefore perform a harder operation: incorporate the new answer while preserving the historical conditions under which the old answer is still true.  \nThe difficulty lies in the locate-and-edit paradigm itself: it locates a factual association (s, r, o) and overwrites it with a new target o′. This works for single-valued facts, but it mismatches temporal knowledge, where the same (s, r) pair can map to different objects under different time contexts. As a result, standard editors often collapse time-conditioned behavior into a single edited answer, degrading historical recall after temporal edits (Yin et al., 2024 ; Cheng et al., 2024 ; Zhao et al., 2026) (Figure 1) . Existing temporal-editing strategies attempt to patch this gap by storing separate parameters per time period (Zhao et al., 2026) or splitting updates into multiple independent edits (Yin et al., 2024) . Yet these approaches leave a key question unanswered: if the unedited model can already answer time-conditioned prompts, why should editing ignore the model’s own temporal computation?  \nWe take a fundamentally different approach. Rather than imposing external temporal structure, we first ask how the model internally resolves temporal ambiguity, and then design an editing method that works with—not against—this mechanism. ","cbCailEnACK89Fyg","https://ap.wps.com/l/cbCailEnACK89Fyg","pdf",1092713,1,16,"English","en",105,"# Abstract\n# Introduction\n# Related Work\n## Locate-then-Edit Knowledge Editing\n## Temporal Editing","[{\"question\":\"What limitation does temporal knowledge editing reveal in the locate-and-edit paradigm?\",\"answer\":\"When a fact changes over time, the update should become current while the old answer may still be correct in historical contexts. The locate-and-edit approach overwrites associations with a single new target, collapsing time-conditioned behavior and harming historical recall.\"},{\"question\":\"How do the authors explain an LLM’s ability to handle temporal answers internally?\",\"answer\":\"Causal tracing shows a two-stage mechanism: early MLP layers retrieve a time-agnostic subject representation, and later layers modulate it with temporal context to yield the time-correct answer.\"},{\"question\":\"What is PRISM Edit and how does it differ from prior temporal editing methods?\",\"answer\":\"PRISM Edit writes a single polysemous representation into the model and relies on the model’s inherent temporal modulation pathway to route predictions correctly for each time context. It avoids architectural modifications, auxiliary modules, and per-period parameter storage.\"}]",1784202433,40,{"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":85,"head_meta":87,"extra_data":89,"updated_unix":27},"prism-edit-one-vector-for-all-temporal-answers","",{"@graph":35,"@context":84},[36,53,67],{"@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/prism-edit-one-vector-for-all-temporal-answers/85315/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":61,"encodingFormat":60,"isAccessibleForFree":62,"interactionStatistic":63},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-16",true,{"@type":64,"interactionType":65,"userInteractionCount":4},"InteractionCounter",{"@type":66},"ViewAction",{"@type":68,"mainEntity":69},"FAQPage",[70,76,80],{"name":71,"@type":72,"acceptedAnswer":73},"What limitation does temporal knowledge editing reveal in the locate-and-edit paradigm?","Question",{"text":74,"@type":75},"When a fact changes over time, the update should become current while the old answer may still be correct in historical contexts. The locate-and-edit approach overwrites associations with a single new target, collapsing time-conditioned behavior and harming historical recall.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How do the authors explain an LLM’s ability to handle temporal answers internally?",{"text":79,"@type":75},"Causal tracing shows a two-stage mechanism: early MLP layers retrieve a time-agnostic subject representation, and later layers modulate it with temporal context to yield the time-correct answer.",{"name":81,"@type":72,"acceptedAnswer":82},"What is PRISM Edit and how does it differ from prior temporal editing methods?",{"text":83,"@type":75},"PRISM Edit writes a single polysemous representation into the model and relies on the model’s inherent temporal modulation pathway to route predictions correctly for each time context. It avoids architectural modifications, auxiliary modules, and per-period parameter storage.","https://schema.org",{"og:url":51,"og:type":86,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":88,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":91},[92,96,100,104,109,114,118,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":93,"show_sort_weight":94,"slug":95},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":97,"show_sort_weight":98,"slug":99},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":101,"show_sort_weight":102,"slug":103},"Exam",70,"exam",{"id":105,"doc_module":4,"doc_module_name":45,"category_name":106,"show_sort_weight":107,"slug":108},5,"Comic",60,"comic",{"id":110,"doc_module":4,"doc_module_name":45,"category_name":111,"show_sort_weight":112,"slug":113},6,"Technology",50,"technology",{"id":115,"doc_module":4,"doc_module_name":45,"category_name":116,"show_sort_weight":28,"slug":117},7,"Healthcare","healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":105,"slug":136},19,"General","general"]