[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-85669-en":3,"doc-seo-85669-105":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},85669,549758252649,"Ivy","https://ap-avatar.wpscdn.com/avatar/8000253669c5317157?_k=1778319167496531819",8,"Research & Report","Faithful Not Corrective Message Format Effects in Multi Hop Agent Relays Are Tier Dependent","Faithful, Not Corrective: Message-Format Effects in Multi-Hop Agent Relays Are Tier-Dependent analyzes whether and how communication structure changes information loss when LLM agents exchange messages across multiple hops. A controlled relay testbed encodes task briefs from twelve atomic facts into five formats over six hops, scoring hop-by-hop fidelity with a fixed strong grader versus ground-truth generation. Results show format effects depend on relay tier: strong relays are nearly lossless under faithful-relay instructions, while structure mainly preserves content without correcting existing errors, and weak relays amplify dispersion across formats.","Faithful, Not Corrective: Message-Format Effects in Multi-Hop Agent Relays Are Tier-Dependent  \nZayx Shawn  \nIndependent Researcher  \narXiv :2607 .09678v 1 [ cs .AI] 12 Jun 2026  \nAbstract  \nWhen LLM agents hand information to one another, does the message format matter? Two literatures give contradictory answers: formatoptimization work reports that structured messages cut cost without hurting accuracy, while format-restriction studies find that imposing structure degrades generation—and neither line has measured what happens when messages traverse multiple hops, where copy fidelity rather than one-shot generation quality dominates. We introduce a controlled relay testbed: task briefs containing twelve programmatically generated atomic facts are re-encoded hop-by-hop in five formats (free natural language, precisioninstructed NL, JSON, triples, key–value) over six hops, scored by a fixed strong grader against programmatic ground truth, across two relaycapability tiers, a cognitive-load condition, anda paired-fork error injection. In short, in this two-tier case study: format effects depend on the relay model—and, at the strong tier, not on how busy the relay is—while structure preserves content without repairing it. Specifically, (i) under faithful-relay instructions a strong relay is nearly lossless: the documented“telephone-game” collapse does not occur, and for all but free NL the residual loss concentratesin the first encoding step; (ii) at the strong tier, adding per-hop cognitive load raises generation cost by 24–53% while format-level fidelity changes are bounded within ±1.8 points (simultaneous bound) . Under a weak (1.5B) relay,(iii) the dispersion of hop-6 recall across formats grows by a factor of 8.7 by QA (2 .2 × judge-free) on matched items (SD across format means, 95% CI 5 .3–15.5; the spread grows from 2.3 to 20.5 points), driven by two opposing mechanisms—an encoding toll paid by both rigid formats and drift resistance associated specifically with the fixed-key JSON schema—that flip the format ranking in transit; and (iv) in a paired-fork injection at the weak tier, a redundancy-free wrong value, once present, persists to the final hop in 83–100% of chains in  \nevery format (surface presence; grader-level 76– 86%), while collateral damage to neighboring facts is undetectable (per-format upper bounds 0–13 points; JSON’s bound is ≤ 0) . Structure buys a faithful, error-localizing channel—not an error-correcting code—and format choice should follow the weakest relay in the pipeline.  \n1 Introduction  \nMulti-agent LLM systems are, at bottom, relay systems. A planner summarizes a task for an executor, the executor hands findings to a reviewer, an orchestrator re-briefs a sub-agent: at every boundary, task-critical information is re-encoded by one model for consumption by another (Hong et al., 2024 ; Qian et al., 2024 ; Du et al., 2024) . Production pipelines increasingly place small models at these boundaries for cost reasons (Narayan et al., 2025 ; Z˙ywot et al., 2026), and empirical failure analyses attribute over a third of multi-agent failures to inter-agent misalignment rather than to basemodel capability (Cemri et al., 2025) . What gets lost at each handoff—and what the message format does to that loss—is therefore a load-bearing design question.  \nTwo adjacent literatures answer it in opposite directions. On one side, format-optimization work shows that letting agents communicate in structured, non-natural-language formats preserves accuracy at a fraction of the token cost (Chen et al., 2024), and protocol and safety guidelines recommend schema-constrained messages specifically to prevent semantic drift (Wibowo and Polyzos, 2025) . On the other, format-restriction studies find the opposite sign: forcing models to generate inside rigid schemas measurably degrades reasoning (Tamet al., 2024), and natural-language tool interfaces outperform structured schemas (Johnson et al., 2025) . Meanwhile, a separate dist","cbCait55c07kYGGS","https://ap.wps.com/l/cbCait55c07kYGGS","pdf",297692,1,11,"English","en",105,"# Introduction\n## Motivation: Contradictory format literatures\n## Distortion across relay chains and the missing experimental cell\n## Controlled relay testbed and evaluation design","[{\"question\":\"Why is message format expected to matter differently in multi-hop relays than in single interactions?\",\"answer\":\"In single interactions, format mainly influences one-shot generation quality. In multi-hop relays, format also affects how well later hops can re-encode content they did not author, so cumulative drift and copy fidelity dominate.\"},{\"question\":\"How does the study measure fidelity across hops?\",\"answer\":\"Task briefs are generated from twelve synthetic atomic facts and used as programmatic ground truth. Fidelity is scored using QA recall via a fixed strong grader at temperature 0 and also via judge-free verbatim string recall.\"},{\"question\":\"What is the key finding about faithful-relay instructions and relay tier?\",\"answer\":\"Under faithful-relay instructions, a strong relay is nearly lossless and avoids the “telephone-game” collapse. Structure preserves content but does not repair errors, and format effects depend on relay model tier rather than relay busyness at the strong tier.\"}]",1784205502,28,{"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},"faithful-not-corrective-message-format-effects-in-multi-hop-agent-relays-are-tier-dependent","",{"@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/faithful-not-corrective-message-format-effects-in-multi-hop-agent-relays-are-tier-dependent/85669/",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},"Why is message format expected to matter differently in multi-hop relays than in single interactions?","Question",{"text":74,"@type":75},"In single interactions, format mainly influences one-shot generation quality. In multi-hop relays, format also affects how well later hops can re-encode content they did not author, so cumulative drift and copy fidelity dominate.","Answer",{"name":77,"@type":72,"acceptedAnswer":78},"How does the study measure fidelity across hops?",{"text":79,"@type":75},"Task briefs are generated from twelve synthetic atomic facts and used as programmatic ground truth. Fidelity is scored using QA recall via a fixed strong grader at temperature 0 and also via judge-free verbatim string recall.",{"name":81,"@type":72,"acceptedAnswer":82},"What is the key finding about faithful-relay instructions and relay tier?",{"text":83,"@type":75},"Under faithful-relay instructions, a strong relay is nearly lossless and avoids the “telephone-game” collapse. Structure preserves content but does not repair errors, and format effects depend on relay model tier rather than relay busyness at the strong tier.","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},[92,96,100,104,109,114,119,122,127,130,134],{"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":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":45,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":45,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":45,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":45,"category_name":136,"show_sort_weight":105,"slug":137},19,"General","general"]