Spaces:
Running
Running
| <html lang="en"> | |
| <head> | |
| <meta charset="UTF-8" /> | |
| <meta name="viewport" content="width=device-width, initial-scale=1.0" /> | |
| <meta name="description" content="SearchGen-Bench model leaderboard with per-component, domain, and failure-mode breakdowns." /> | |
| <title>SearchGen-Bench</title> | |
| <link rel="preconnect" href="https://fonts.googleapis.com"> | |
| <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin> | |
| <link | |
| href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600&family=Outfit:wght@300;400;500;600;700&display=swap" | |
| rel="stylesheet"> | |
| <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.2/css/all.min.css"> | |
| <link rel="stylesheet" href="./style.css"> | |
| </head> | |
| <body> | |
| <div class="bg-blobs"> | |
| <div class="blob blob-1"></div> | |
| <div class="blob blob-2"></div> | |
| <div class="blob blob-3"></div> | |
| </div> | |
| <header> | |
| <h1>SearchGen-Bench</h1> | |
| <p class="subtitle">Evaluating search-grounded image generation across knowledge boundaries</p> | |
| <div class="stat-grid" aria-label="Benchmark summary"> | |
| <article><strong>751</strong><span>evaluation prompts</span></article> | |
| <article><strong>22</strong><span>knowledge domains</span></article> | |
| <article><strong>12</strong><span>failure modes</span></article> | |
| <article><strong>10</strong><span>score components</span></article> | |
| </div> | |
| <div id="intro-container" class="intro-container"> | |
| <div id="intro-markdown" class="markdown-body"></div> | |
| </div> | |
| <div class="expandable-controls"> | |
| <a href="mailto:jasper.whz@outlook.com" class="expand-btn link-btn"> | |
| <i class="fa-solid fa-rocket"></i> Submit Here | |
| </a> | |
| <a href="https://arxiv.org/abs/2607.05382" class="expand-btn link-btn" target="_blank" rel="noopener"> | |
| <svg viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg"> | |
| <path d="M19 3L5 21" stroke="#B31B1B" stroke-width="4" stroke-linecap="round" /> | |
| <path d="M5 3L19 21" stroke="#5f6368" stroke-width="4" stroke-linecap="round" opacity="0.6" /> | |
| </svg> Paper | |
| </a> | |
| <a href="https://haozheh3.github.io/SearchGen/" class="expand-btn link-btn" target="_blank" rel="noopener"> | |
| <i class="fa-solid fa-globe"></i> Project Page | |
| </a> | |
| <a href="https://huggingface.co/papers/2607.05382" class="expand-btn link-btn" target="_blank" rel="noopener"> | |
| <span class="icon-emoji">🤗</span> HF Paper | |
| </a> | |
| <a href="https://github.com/HaozheH3/SearchGen" class="expand-btn link-btn" target="_blank" rel="noopener" aria-label="Code repository"> | |
| <i class="fa-brands fa-github"></i> Code | |
| </a> | |
| <a href="https://huggingface.co/datasets/JasperHaozhe/SearchGen-Bench" class="expand-btn link-btn" target="_blank" rel="noopener"> | |
| <span class="icon-emoji">🤗</span> Data | |
| </a> | |
| </div> | |
| </header> | |
| <script src="https://cdn.jsdelivr.net/npm/marked/marked.min.js"></script> | |
| <main> | |
| <div class="controls"> | |
| <div class="search-container"> | |
| <input type="text" id="model-search" placeholder="Search models..." aria-label="Search models"> | |
| </div> | |
| <div class="select-container"> | |
| <label for="board-select">Scoreboard</label> | |
| <select id="board-select"> | |
| <option value="components">Components</option> | |
| <option value="domains">Domains</option> | |
| <option value="failure_modes">Failure modes</option> | |
| </select> | |
| </div> | |
| <div class="select-container" id="stratum-control"> | |
| <label for="stratum-select">Prompt set</label> | |
| <select id="stratum-select"> | |
| <option value="All">All 751</option> | |
| <option value="NoSearch">NoSearch</option> | |
| <option value="SearchIntensive">SearchIntensive</option> | |
| <option value="VisualSearch">VisualSearch</option> | |
| <option value="TextualSearch">TextualSearch</option> | |
| </select> | |
| </div> | |
| <div id="stats-summary"> | |
| <!-- Stats will be injected here --> | |
| </div> | |
| </div> | |
| <div id="top-scrollbar-wrapper" class="scrollbar-wrapper"> | |
| <div id="top-scrollbar-content"></div> | |
| </div> | |
| <div class="table-outer-container"> | |
| <div class="table-container" id="bottom-scrollbar-wrapper"> | |
| <table id="leaderboard-table"> | |
| <thead> | |
| <tr id="table-header"> | |
| <!-- Headers will be injected here --> | |
| </tr> | |
| </thead> | |
| <tbody id="table-body"> | |
| <!-- Rows will be injected here --> | |
| <tr> | |
| <td colspan="100%" style="text-align: center; padding: 4rem;">Loading benchmark data...</td> | |
| </tr> | |
| </tbody> | |
| </table> | |
| </div> | |
| </div> | |
| <p class="table-note">Click any column header to sort. Domains and failure modes are multi-label: group counts overlap. Coverage is shown because missing results are zero-filled or excluded according to the canonical model-specific policy.</p> | |
| <section id="domain-glossary" class="taxonomy-panel" hidden> | |
| <div class="taxonomy-heading"> | |
| <div><span class="eyebrow">Reference</span><h3>Domain abbreviations</h3></div> | |
| <p>Twenty-two content domains represented in SearchGen-Bench.</p> | |
| </div> | |
| <div class="glossary-grid"> | |
| <div><code>ABS</code><span><b>Abstract</b> — Abstract concepts, emotions, and symbolism.</span></div> | |
| <div><code>ARC</code><span><b>Architecture</b> — Buildings, structures, interiors, and urban spaces.</span></div> | |
| <div><code>ART</code><span><b>Art & Design</b> — Artistic styles, illustration, visual design, and aesthetics.</span></div> | |
| <div><code>COM</code><span><b>Commercial</b> — Advertising, branding, marketing, and product imagery.</span></div> | |
| <div><code>CUL</code><span><b>Culture</b> — Cultural artifacts, traditions, heritage, and customs.</span></div> | |
| <div><code>EDU</code><span><b>Education</b> — Educational diagrams and instructional content.</span></div> | |
| <div><code>FAS</code><span><b>Fashion</b> — Apparel, accessories, style, and textiles.</span></div> | |
| <div><code>FOO</code><span><b>Food</b> — Cuisine, dishes, ingredients, and culinary arts.</span></div> | |
| <div><code>GAM</code><span><b>Gaming</b> — Video games, virtual worlds, and game characters.</span></div> | |
| <div><code>HIS</code><span><b>History</b> — Historical events, figures, periods, and artifacts.</span></div> | |
| <div><code>INF</code><span><b>Information Design</b> — Data visualization, infographics, and diagrams.</span></div> | |
| <div><code>MED</code><span><b>Medical</b> — Anatomy, medical illustration, and health.</span></div> | |
| <div><code>MIL</code><span><b>Military</b> — Weapons, military equipment, and warfare.</span></div> | |
| <div><code>MUS</code><span><b>Music</b> — Instruments, performances, and music-related imagery.</span></div> | |
| <div><code>OBJ</code><span><b>Objects</b> — Products, everyday objects, and artifacts.</span></div> | |
| <div><code>OTH</code><span><b>Other</b> — Miscellaneous or uncategorized content.</span></div> | |
| <div><code>PEO</code><span><b>People</b> — Portraits, human figures, and fashion on people.</span></div> | |
| <div><code>PSY</code><span><b>Psychology</b> — Mental states and psychological concepts.</span></div> | |
| <div><code>SCI</code><span><b>Science & Nature</b> — Biology, physics, natural phenomena, and scientific illustration.</span></div> | |
| <div><code>SPO</code><span><b>Sports</b> — Athletics, sports, fitness, and recreation.</span></div> | |
| <div><code>TEC</code><span><b>Technology</b> — Devices, interfaces, software, and technical diagrams.</span></div> | |
| <div><code>TXT</code><span><b>Text & Typography</b> — Text-heavy content, calligraphy, typography, and signage.</span></div> | |
| </div> | |
| </section> | |
| <section id="failure-glossary" class="taxonomy-panel" hidden> | |
| <div class="taxonomy-heading"> | |
| <div><span class="eyebrow">Reference</span><h3>Failure-mode abbreviations</h3></div> | |
| <p>Twelve recurring categories distilled from real production requests.</p> | |
| </div> | |
| <div class="glossary-grid"> | |
| <div><code>TK-R</code><span><b>Temporal — Recent</b> — Newly emerged or recently changed entities, events, products, and trends.</span></div> | |
| <div><code>TK-C</code><span><b>Temporal — Current</b> — Up-to-date states such as rankings, live events, locations, or social conditions.</span></div> | |
| <div><code>EIK</code><span><b>Entity & IP Knowledge</b> — Accurate identity and appearance of named people, characters, brands, products, or landmarks.</span></div> | |
| <div><code>CSV</code><span><b>Concept & Symbol Visualization</b> — Precise concepts, flags, symbols, and diagrams requiring exact representation.</span></div> | |
| <div><code>FHA</code><span><b>Factual & Historical Accuracy</b> — Correct facts, periods, artifacts, scientific details, and historical depictions.</span></div> | |
| <div><code>CS</code><span><b>Cultural Specificity</b> — Authentic culture-bound practices, aesthetics, artifacts, and conventions.</span></div> | |
| <div><code>VUR</code><span><b>Visual / UI / UX</b> — Specific interfaces, layouts, rendering conventions, spatial logic, and visual relationships.</span></div> | |
| <div><code>DVZ</code><span><b>Data Visualization</b> — Accurate charts, graphs, timelines, diagrams, and infographics.</span></div> | |
| <div><code>THG</code><span><b>Text / Typography</b> — Correct, readable, and stylistically appropriate text rendered inside images.</span></div> | |
| <div><code>CCS</code><span><b>Complex Composite</b> — Coherent multi-element scenes, spatial arrangements, and cross-concept compositions.</span></div> | |
| <div><code>VUA</code><span><b>Vague / Abstract</b> — Visualizing underspecified ideas, moods, metaphors, and abstract concepts.</span></div> | |
| <div><code>IRR</code><span><b>Implicit Reasoning</b> — Requirements that must be inferred rather than stated directly, including contextual and relational constraints.</span></div> | |
| </div> | |
| </section> | |
| <!-- Citation Section --> | |
| <section id="citation-section" class="citation-section"> | |
| <h3>Cite SearchGen-Bench</h3> | |
| <p>If you found SearchGen-Bench useful, please cite us at:</p> | |
| <div class="citation-block-wrapper"> | |
| <button id="copy-citation-btn" class="copy-btn"> | |
| <i class="fa-regular fa-copy"></i> Copy | |
| </button> | |
| <pre id="citation-content"></pre> | |
| </div> | |
| </section> | |
| </main> | |
| <script type="module" src="/main.js"></script> | |
| </body> | |
| </html> | |