Technical report PDF renamed and stamped with its author (2026-11-02)

.gitattributes

@@ -44,3 +44,4 @@ graphics/home/research-adaptation-v1.jpg filter=lfs diff=lfs merge=lfs -text
 graphics/home/hero-understanding-v1.jpg filter=lfs diff=lfs merge=lfs -text
 graphics/hero-understanding-v1.png filter=lfs diff=lfs merge=lfs -text
 publications/Lorm-Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf filter=lfs diff=lfs merge=lfs -text
+publications/Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf filter=lfs diff=lfs merge=lfs -text

manifest.json

@@ -23,16 +23,16 @@
     "benchmark/src/pages/leaderboard.html": "6ba4907bda377904e005572c40969d3167b12d54a4899659fa50dde9f5ddf76f",
     "benchmark/src/pages/method.html": "9981ac808f6eae01f428ded9eca962d0bf3228a70a655c9ab8a99482294954c4",
     "benchmark/src/pages/reliability.html": "e48f1584620287b8286c3d68fcd1a82763fd43e4e1fe1bd725ef89259f97e2fc",
-    "build.py": "c6abf91ec9a17da2fc796420bc3771c396b4a71753dff76ae56d80baabcce7d9",
+    "build.py": "24e637733a4f113c480b2d35b54837bb3a8a0d0a1e68e0e3c39360deaa45eb54",
     "data/explorer-release.json": "085e2a3cbdf5427055427abebf2f92c453805f82257cdbb26e2b894638612e97",
     "data/legal-copy.json": "66458ad0e70b6be38a49ac698d14a2150ce5d5661593c03d68f158df957b336f",
     "data/licenses/CC-BY-4.0.txt": "9ba9550ad48438d0836ddab3da480b3b69ffa0aac7b7878b5a0039e7ab429411",
     "data/preview.schema.json": "dc0262257a6055aa9c4a3856eba0a3fa6694c4748fbff591b7034cf5b7e278c7",
     "data/publication-release.json": "7b4d40bca16ec7843a59bd0b0f13ac52f5b78c30eec512e15e2c34e9ab6bc674",
     "legal_copy.py": "efcd22fcfa16af1219fe724751954535f824027b46e9f12823f4cdadcf4213da",
-    "paper_beta.py": "d860be846e5fe3240aef14d56d36eb0e594b5432b5849e35258a239517af690a",
+    "paper_beta.py": "7984d69f86bb0ab13918de02e191eb4aa7fec7de6c5ad3b9ef20756fb6f76fb8",
     "publication.py": "06a9c6f67242c1118120ed31abf7dbc7fa4fb6d3224bf9e62da0d19a7efc5b4b",
-    "render_paper_beta_pdf.py": "81e67170e1d4e97509b8d9857df605ff43e3d80ab6380ec72e9b79fbe4ea199f",
+    "render_paper_beta_pdf.py": "4135fa359c9938c4b478875a52697c750122e3a1be983a28fdf9a08b037c5dec",
     "render_publication_figures.py": "6cc13bd836bd95da9fbee23d1b6de4cf17cce359bc65db28a5a89eaa85b2e2ed",
     "sensus.py": "deff2e2631b277d3a56292796ca5e68e480e75c897f6e0df1c2c55612df07b18",
     "sensus_data.py": "b25cbfda49711b1e3c32e959adc6064f26b7a59bcc12c7878355e445526bb8c9",
@@ -81,7 +81,7 @@
     "src/pages/research.html": "bf493c6ea645c95727684eed2fe9e9e81307d9fff11bd40241c501af065c54e2",
     "src/pages/sensus-v1.html": "1e45c472373cd4f2cc9ce3270817ff260aa68ceebfd86ff7ce2229e21ecae388",
     "src/pages/training.html": "d5f77262791c5d0a7b78ef6e679b32604dc7fa82aec93d1496b9053b8b641773",
-    "src/publications/Lorm-Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf": "6152a0e672145db4de496eca5e7d80cebadd21ae4458bfb46345c80d31bc12ef"
+    "src/publications/Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf": "fe196d71235153ef0f4a8555055f6673dde72c19f7559c40893d42b2014d526d"
   },
   "imported_sources": {
     "web/lorm-theme.css": "7b4653376ad076a13bc67174bff2a9f92dca529f7bdd57fd3172bb705c94d90b",
@@ -189,10 +189,10 @@
     "graphics/research-discovery-v1.webp": "978cb82346d9fa09eaf39d7c966bbb9d235035ae57174f038e9d990c275abd63",
     "graphics/research-reasoning-v1.webp": "0f4be06aca754f6096d79ef6b7ee538ad11684b8635cfa5c2645e4a258ac00c3",
     "index.html": "94523f6091d66cda75b49c9662ccdfc0b24e49b137660bd4d50193fa2a0227f9",
-    "publications/Lorm-Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf": "6152a0e672145db4de496eca5e7d80cebadd21ae4458bfb46345c80d31bc12ef",
-    "redirects.json": "702a551b2ab9995ba4d20a2fe13eac1db5f50ed900bb2bba7e0eeacfa2908ec4",
-    "research/index.html": "135ac7a7308ef2e94b6b73bee3e83f71b0bc934138211a44f439bdfd87ee3049",
-    "research/teaching-models-to-read-learner-reasoning/index.html": "4b7995f7ad8bb88a977647316b2fc524d61692fe63c5e13c5d24ca7848c5ec79",
+    "publications/Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf": "fe196d71235153ef0f4a8555055f6673dde72c19f7559c40893d42b2014d526d",
+    "redirects.json": "339a40dbbeee2b158b5f54ac725739dcc76b9d0a306b3421984d3be7d08c4530",
+    "research/index.html": "c9926131b6e83f774eeeff414fd79049bab774bd3916299db23c994cb72cfa88",
+    "research/teaching-models-to-read-learner-reasoning/index.html": "1ab69dfc19732abaefadcac186eac7a3dee66c4f002bac352631184fd0a0253b",
     "sitemap.xml": "033b3222c3ff90404bb0e2b5662912f276224422f2dc406c2c7d816e5454e513",
     "training/index.html": "58364404b8c0f1f986de820ca7504995e34af9cfed4d242cabc0a489b21902c1"
   }

publications/{Lorm-Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf → Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6152a0e672145db4de496eca5e7d80cebadd21ae4458bfb46345c80d31bc12ef
-size 7812091
+oid sha256:fe196d71235153ef0f4a8555055f6673dde72c19f7559c40893d42b2014d526d
+size 7812799

research/index.html

@@ -9,7 +9,7 @@
 <li><div class="pub-meta"><span class="type">Paper</span><time datetime="2026-11-02">November 2, 2026</time></div>
 <div class="pub-body"><h3><a href="teaching-models-to-read-learner-reasoning/index.html">Teaching Models to Read and Respond to a Learner's Reasoning: SENSUS Bench and an Open SFT Dataset for Grade 4 and 5 Fraction and Decimal Errors</a></h3>
 <p>Fine-tuning an open 27B model on 1,015 synthetic conversations raises its checklist score from 0.290 before training to 0.825 on average over two runs, ahead of every general-purpose model we tested. The report sets out the benchmark, the data, the grading checks and the limits.</p>
-<p class="card-links"><a href="teaching-models-to-read-learner-reasoning/index.html">Read</a><a href="../publications/Lorm-Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf" download>PDF</a><a href="../benchmarks/sensus-bench/index.html">Benchmark</a><a href="../training/index.html">Data</a></p></div></li>
+<p class="card-links"><a href="teaching-models-to-read-learner-reasoning/index.html">Read</a><a href="../publications/Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf" download>PDF</a><a href="../benchmarks/sensus-bench/index.html">Benchmark</a><a href="../training/index.html">Data</a></p></div></li>
 </ul>
 </section>
 </main><footer class="site-footer"><div class="footer-inner"><div class="footer-brand"><a class="brand" href="../index.html" aria-label="Lorm home"><span class="wordmark">Lorm<span class="wordmark-dot">.</span></span></a><p>Turning advances in AI into advances in human understanding.</p><address>315 W 35th Street<br>New York, NY 10001</address></div><nav class="footer-col" aria-label="Site"><h2>Lorm</h2><a href="index.html">Research</a><a href="../benchmarks/index.html">Benchmarks</a><a href="../datasets/index.html">Data</a><a href="../about/index.html">About</a></nav><nav class="footer-col" aria-label="Benchmarks"><h2>Benchmarks</h2><a href="../benchmarks/index.html">All Benchmarks</a><a href="../benchmarks/sensus-bench/index.html">SENSUS Bench v1.0</a></nav><div class="footer-legal"><p>© 2026 Lorm Labs, Inc.</p><nav class="footer-links" aria-label="Legal"><a href="#terms-of-service" data-dialog-open="terms-of-service">Terms of Service</a><a href="#privacy-policy" data-dialog-open="privacy-policy">Privacy Policy</a><a href="https://lorm.ai/contact" target="_blank" rel="noopener" data-dialog-open="contact-us">Contact Us</a></nav></div></div></footer><dialog id="contact-us" class="site-dialog" aria-labelledby="contact-title"><button type="button" class="dialog-close" data-dialog-close aria-label="Close contact window">×</button><a class="dialog-fallback-close" href="#main">Close window</a><h2 id="contact-title">Contact Us</h2>

research/teaching-models-to-read-learner-reasoning/index.html

@@ -1,4 +1,4 @@
-<!doctype html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><meta name="robots" content="noindex,nofollow"><meta name="referrer" content="same-origin"><meta http-equiv="Content-Security-Policy" content="default-src 'none'; style-src 'self'; script-src 'self'; img-src 'self' data:; font-src 'self'; connect-src 'self'; object-src 'none'; base-uri 'none'; form-action 'self'"><meta name="theme-color" content="#2C5672"><title>Teaching Models to Read and Respond to a Learner&#x27;s Reasoning | Lorm</title><meta name="description" content="A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result."><link rel="canonical" href="https://lorm.ai/research/teaching-models-to-read-learner-reasoning"><meta property="og:type" content="article"><meta property="og:site_name" content="Lorm"><meta property="og:locale" content="en_US"><meta property="og:title" content="Teaching Models to Read and Respond to a Learner&#x27;s Reasoning | Lorm"><meta property="og:description" content="A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result."><meta property="og:url" content="https://lorm.ai/research/teaching-models-to-read-learner-reasoning"><meta property="og:image" content="https://lorm.ai/assets/social/sensus-v1.png"><meta property="og:image:width" content="1200"><meta property="og:image:height" content="630"><meta property="og:image:alt" content="SENSUS Bench v1.0. Explore the leaderboard, play Duel and get the dataset."><meta name="twitter:card" content="summary_large_image"><meta name="twitter:title" content="Teaching Models to Read and Respond to a Learner&#x27;s Reasoning | Lorm"><meta name="twitter:description" content="A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result."><meta name="twitter:image" content="https://lorm.ai/assets/social/sensus-v1.png"><meta name="twitter:image:alt" content="SENSUS Bench v1.0. Explore the leaderboard, play Duel and get the dataset."><script type="application/ld+json">{"@context":"https://schema.org","@graph":[{"@type":"ScholarlyArticle","headline":"Teaching Models to Read and Respond to a Learner's Reasoning","description":"A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result.","url":"https://lorm.ai/research/teaching-models-to-read-learner-reasoning","datePublished":"2026-11-02","author":{"@type":"Person","name":"Karen Wang","affiliation":{"@type":"Organization","name":"Georgetown University"}},"publisher":{"@type":"Organization","name":"Lorm Labs","alternateName":"Lorm","url":"https://lorm.ai/","address":{"@type":"PostalAddress","streetAddress":"315 W 35th Street","addressLocality":"New York","addressRegion":"NY","postalCode":"10001","addressCountry":"US"}},"isAccessibleForFree":true,"inLanguage":"en"},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://lorm.ai/"},{"@type":"ListItem","position":2,"name":"Research","item":"https://lorm.ai/research"},{"@type":"ListItem","position":3,"name":"Teaching Models to Read and Respond to a Learner's Reasoning","item":"https://lorm.ai/research/teaching-models-to-read-learner-reasoning"}]}]}</script><link rel="preload" href="../../assets/fonts/Montserrat-Variable.ttf" as="font" type="font/ttf" crossorigin><link rel="stylesheet" href="../../assets/style.css"><script src="../../assets/site.js" defer></script><script src="../../assets/lorm.js" defer></script><link rel="stylesheet" href="../../assets/publication-reading.css"><script src="../../benchmarks/sensus-bench/assets/pilot-math.js" defer></script><script src="../../assets/publication.js" defer></script><link rel="stylesheet" href="../../assets/paper-print.css" media="print"><script src="../../assets/contact.js" defer></script></head><body><a class="skip" href="#main">Skip to content</a><header class="site-header"><div class="header-inner"><a class="brand" href="../../index.html" aria-label="Lorm home"><span class="wordmark">Lorm<span class="wordmark-dot">.</span></span></a><nav class="main-nav" aria-label="Main"><div class="nav-item"><a class="nav-top" href="../index.html">Research</a></div><div class="nav-item"><a class="nav-top" href="../../benchmarks/index.html" aria-current="location">Benchmarks<span class="nav-caret" aria-hidden="true"></span></a><div class="nav-menu"><a class="menu-link" href="../../benchmarks/index.html"><span class="menu-title">All Benchmarks</span></a><div class="menu-group"><p class="menu-group-title">SENSUS Bench v1.0</p><div class="menu-sub"><a href="../../benchmarks/sensus-bench/index.html">Results</a><a href="index.html">Paper</a><a href="../../benchmarks/sensus-bench/duel/index.html">Duel</a><a href="../../training/index.html">Data</a><a href="../../benchmarks/sensus-bench/families/index.html">Evaluation Cases</a><a href="../../benchmarks/sensus-bench/method/index.html">Method</a></div></div></div></div><div class="nav-item"><a class="nav-top" href="../../datasets/index.html">Data</a></div><div class="nav-item"><a class="nav-top" href="../../about/index.html">About</a></div></nav><details class="mobile-menu"><summary>Menu</summary><nav aria-label="Mobile main"><div class="mobile-group"><a class="mobile-single" href="../index.html">Research</a></div><div class="mobile-group"><p class="mobile-head">Benchmarks</p><a href="../../benchmarks/index.html">All Benchmarks</a><p class="mobile-subhead">SENSUS Bench v1.0</p><a class="mobile-sub" href="../../benchmarks/sensus-bench/index.html">Results</a><a class="mobile-sub" href="index.html">Paper</a><a class="mobile-sub" href="../../benchmarks/sensus-bench/duel/index.html">Duel</a><a class="mobile-sub" href="../../training/index.html">Data</a><a class="mobile-sub" href="../../benchmarks/sensus-bench/families/index.html">Evaluation Cases</a><a class="mobile-sub" href="../../benchmarks/sensus-bench/method/index.html">Method</a></div><div class="mobile-group"><a class="mobile-single" href="../../datasets/index.html">Data</a></div><div class="mobile-group"><a class="mobile-single" href="../../about/index.html">About</a></div></nav></details></div></header><main id="main" tabindex="-1"><nav class="breadcrumbs" aria-label="Breadcrumb"><ol><li><a href="../../benchmarks/index.html">Benchmarks</a></li><li><a href="../../benchmarks/sensus-bench/index.html">SENSUS Bench v1.0</a></li><li><span aria-current="page">Paper</span></li></ol></nav><nav class="flow-steps" aria-label="SENSUS Bench v1.0"><span class="flow-steps-title">SENSUS Bench v1.0</span><ol><li><a href="../../benchmarks/sensus-bench/index.html"><span class="flow-step-number">1</span>Results</a></li><li><a href="index.html" aria-current="step"><span class="flow-step-number">2</span>Paper</a></li><li><a href="../../benchmarks/sensus-bench/duel/index.html"><span class="flow-step-number">3</span>Duel</a></li><li><a href="../../training/index.html"><span class="flow-step-number">4</span>Data</a></li></ol></nav><section class="page-head publication-head paper-head"><p class="kicker">Research paper · technical report</p><h1>Teaching Models to Read and Respond to a Learner's Reasoning</h1><p>SENSUS Bench and an Open SFT Dataset for Grade 4 and 5 Fraction and Decimal Errors</p><p class="publication-meta">Karen Wang · Georgetown University · November 2, 2026</p><div class="actions"><a class="button primary" href="../../publications/Lorm-Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf" download>Download PDF</a></div></section>
+<!doctype html><html lang="en"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><meta name="robots" content="noindex,nofollow"><meta name="referrer" content="same-origin"><meta http-equiv="Content-Security-Policy" content="default-src 'none'; style-src 'self'; script-src 'self'; img-src 'self' data:; font-src 'self'; connect-src 'self'; object-src 'none'; base-uri 'none'; form-action 'self'"><meta name="theme-color" content="#2C5672"><title>Teaching Models to Read and Respond to a Learner&#x27;s Reasoning | Lorm</title><meta name="description" content="A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result."><link rel="canonical" href="https://lorm.ai/research/teaching-models-to-read-learner-reasoning"><meta property="og:type" content="article"><meta property="og:site_name" content="Lorm"><meta property="og:locale" content="en_US"><meta property="og:title" content="Teaching Models to Read and Respond to a Learner&#x27;s Reasoning | Lorm"><meta property="og:description" content="A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result."><meta property="og:url" content="https://lorm.ai/research/teaching-models-to-read-learner-reasoning"><meta property="og:image" content="https://lorm.ai/assets/social/sensus-v1.png"><meta property="og:image:width" content="1200"><meta property="og:image:height" content="630"><meta property="og:image:alt" content="SENSUS Bench v1.0. Explore the leaderboard, play Duel and get the dataset."><meta name="twitter:card" content="summary_large_image"><meta name="twitter:title" content="Teaching Models to Read and Respond to a Learner&#x27;s Reasoning | Lorm"><meta name="twitter:description" content="A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result."><meta name="twitter:image" content="https://lorm.ai/assets/social/sensus-v1.png"><meta name="twitter:image:alt" content="SENSUS Bench v1.0. Explore the leaderboard, play Duel and get the dataset."><script type="application/ld+json">{"@context":"https://schema.org","@graph":[{"@type":"ScholarlyArticle","headline":"Teaching Models to Read and Respond to a Learner's Reasoning","description":"A technical report introducing SENSUS Bench and an open SFT dataset for grade 4 and 5 fraction and decimal errors, with every recorded result.","url":"https://lorm.ai/research/teaching-models-to-read-learner-reasoning","datePublished":"2026-11-02","author":{"@type":"Person","name":"Karen Wang","affiliation":{"@type":"Organization","name":"Georgetown University"}},"publisher":{"@type":"Organization","name":"Lorm Labs","alternateName":"Lorm","url":"https://lorm.ai/","address":{"@type":"PostalAddress","streetAddress":"315 W 35th Street","addressLocality":"New York","addressRegion":"NY","postalCode":"10001","addressCountry":"US"}},"isAccessibleForFree":true,"inLanguage":"en"},{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://lorm.ai/"},{"@type":"ListItem","position":2,"name":"Research","item":"https://lorm.ai/research"},{"@type":"ListItem","position":3,"name":"Teaching Models to Read and Respond to a Learner's Reasoning","item":"https://lorm.ai/research/teaching-models-to-read-learner-reasoning"}]}]}</script><link rel="preload" href="../../assets/fonts/Montserrat-Variable.ttf" as="font" type="font/ttf" crossorigin><link rel="stylesheet" href="../../assets/style.css"><script src="../../assets/site.js" defer></script><script src="../../assets/lorm.js" defer></script><link rel="stylesheet" href="../../assets/publication-reading.css"><script src="../../benchmarks/sensus-bench/assets/pilot-math.js" defer></script><script src="../../assets/publication.js" defer></script><link rel="stylesheet" href="../../assets/paper-print.css" media="print"><script src="../../assets/contact.js" defer></script></head><body><a class="skip" href="#main">Skip to content</a><header class="site-header"><div class="header-inner"><a class="brand" href="../../index.html" aria-label="Lorm home"><span class="wordmark">Lorm<span class="wordmark-dot">.</span></span></a><nav class="main-nav" aria-label="Main"><div class="nav-item"><a class="nav-top" href="../index.html">Research</a></div><div class="nav-item"><a class="nav-top" href="../../benchmarks/index.html" aria-current="location">Benchmarks<span class="nav-caret" aria-hidden="true"></span></a><div class="nav-menu"><a class="menu-link" href="../../benchmarks/index.html"><span class="menu-title">All Benchmarks</span></a><div class="menu-group"><p class="menu-group-title">SENSUS Bench v1.0</p><div class="menu-sub"><a href="../../benchmarks/sensus-bench/index.html">Results</a><a href="index.html">Paper</a><a href="../../benchmarks/sensus-bench/duel/index.html">Duel</a><a href="../../training/index.html">Data</a><a href="../../benchmarks/sensus-bench/families/index.html">Evaluation Cases</a><a href="../../benchmarks/sensus-bench/method/index.html">Method</a></div></div></div></div><div class="nav-item"><a class="nav-top" href="../../datasets/index.html">Data</a></div><div class="nav-item"><a class="nav-top" href="../../about/index.html">About</a></div></nav><details class="mobile-menu"><summary>Menu</summary><nav aria-label="Mobile main"><div class="mobile-group"><a class="mobile-single" href="../index.html">Research</a></div><div class="mobile-group"><p class="mobile-head">Benchmarks</p><a href="../../benchmarks/index.html">All Benchmarks</a><p class="mobile-subhead">SENSUS Bench v1.0</p><a class="mobile-sub" href="../../benchmarks/sensus-bench/index.html">Results</a><a class="mobile-sub" href="index.html">Paper</a><a class="mobile-sub" href="../../benchmarks/sensus-bench/duel/index.html">Duel</a><a class="mobile-sub" href="../../training/index.html">Data</a><a class="mobile-sub" href="../../benchmarks/sensus-bench/families/index.html">Evaluation Cases</a><a class="mobile-sub" href="../../benchmarks/sensus-bench/method/index.html">Method</a></div><div class="mobile-group"><a class="mobile-single" href="../../datasets/index.html">Data</a></div><div class="mobile-group"><a class="mobile-single" href="../../about/index.html">About</a></div></nav></details></div></header><main id="main" tabindex="-1"><nav class="breadcrumbs" aria-label="Breadcrumb"><ol><li><a href="../../benchmarks/index.html">Benchmarks</a></li><li><a href="../../benchmarks/sensus-bench/index.html">SENSUS Bench v1.0</a></li><li><span aria-current="page">Paper</span></li></ol></nav><nav class="flow-steps" aria-label="SENSUS Bench v1.0"><span class="flow-steps-title">SENSUS Bench v1.0</span><ol><li><a href="../../benchmarks/sensus-bench/index.html"><span class="flow-step-number">1</span>Results</a></li><li><a href="index.html" aria-current="step"><span class="flow-step-number">2</span>Paper</a></li><li><a href="../../benchmarks/sensus-bench/duel/index.html"><span class="flow-step-number">3</span>Duel</a></li><li><a href="../../training/index.html"><span class="flow-step-number">4</span>Data</a></li></ol></nav><section class="page-head publication-head paper-head"><p class="kicker">Research paper · technical report</p><h1>Teaching Models to Read and Respond to a Learner's Reasoning</h1><p>SENSUS Bench and an Open SFT Dataset for Grade 4 and 5 Fraction and Decimal Errors</p><p class="publication-meta">Karen Wang · Georgetown University · November 2, 2026</p><div class="actions"><a class="button primary" href="../../publications/Teaching-Models-to-Read-and-Respond-to-a-Learners-Reasoning-Technical-Report-2026-11-02.pdf" download>Download PDF</a></div></section>
 <div class="publication-layout"><aside class="publication-toc"><button class="publication-toc-toggle" type="button" aria-expanded="false" aria-controls="beta-sections" hidden>On this page</button><nav id="beta-sections" aria-label="Paper sections"><a href="#abstract">Abstract</a><a href="#introduction">1 Introduction</a><a href="#related">2 Related work</a><a href="#benchmark">3 Benchmark</a><a href="#training">4 Data and training</a><a href="#protocol">5 Evaluation setup</a><a href="#results">6 Results</a><a href="#gain">7 What the gain consists of</a><a href="#validity">8 Measurement checks</a><a href="#errors">9 Failure analysis</a><a href="#limitations">10 Limitations</a><a href="#ethics">11 Ethics</a><a href="#release">12 Release</a><a href="#conclusion">13 Conclusion</a><a href="#references">References</a><a href="#appendix">Appendices</a></nav></aside><div class="publication-content">
 
 <section class="section" id="abstract"><h2>Abstract</h2>