| <!DOCTYPE html> |
| <html lang="en"> |
| <head> |
| <meta charset="utf-8" /> |
| <meta name="viewport" content="width=device-width, initial-scale=1" /> |
| <title>Dali — the open verification layer for AI</title> |
| <meta |
| name="description" |
| content="Dali is the open verification layer for AI: it creates, scores, and preserves evidence so AI-assisted outputs can be independently verified, exchanged, and replayed." |
| /> |
| <meta name="color-scheme" content="light dark" /> |
| <style> |
| body { |
| margin: 0; |
| padding: 1.5rem 1rem 2rem; |
| font-family: system-ui, -apple-system, sans-serif; |
| line-height: 1.55; |
| color: CanvasText; |
| background: Canvas; |
| } |
| |
| main { |
| max-width: 42rem; |
| margin: 0 auto; |
| } |
| |
| h1 { |
| margin: 0 0 0.25rem; |
| font-size: 1.85rem; |
| letter-spacing: 0.04em; |
| } |
| |
| h3 { |
| margin: 0 0 0.35rem; |
| font-size: 1rem; |
| font-weight: 600; |
| } |
| |
| .lede { |
| margin: 0 0 1.25rem; |
| opacity: 0.85; |
| } |
| |
| hr { |
| border: 0; |
| border-top: 1px solid color-mix(in srgb, CanvasText 18%, transparent); |
| margin: 1.25rem 0; |
| } |
| |
| h2 { |
| font-size: 1rem; |
| margin: 0 0 0.85rem; |
| } |
| |
| .artifact { |
| margin: 0 0 1rem; |
| } |
| |
| .artifact strong a { |
| color: CanvasText; |
| text-decoration: none; |
| } |
| |
| .artifact strong a:hover { |
| text-decoration: underline; |
| } |
| |
| .artifact p { |
| margin: 0.15rem 0 0; |
| opacity: 0.8; |
| font-size: 0.95rem; |
| } |
| |
| .status { |
| font-style: italic; |
| opacity: 0.7; |
| font-size: 0.85rem; |
| } |
| |
| pre { |
| margin: 0 0 1rem; |
| padding: 0.85rem 1rem; |
| overflow-x: auto; |
| font-family: ui-monospace, Menlo, Consolas, monospace; |
| font-size: 0.82rem; |
| line-height: 1.45; |
| border: 1px solid color-mix(in srgb, CanvasText 18%, transparent); |
| background: color-mix(in srgb, CanvasText 4%, Canvas); |
| } |
| |
| .cta { |
| margin: 0; |
| font-weight: 600; |
| } |
| |
| .cta a { |
| color: LinkText; |
| text-decoration: none; |
| } |
| |
| .cta a:hover { |
| text-decoration: underline; |
| } |
| |
| table { |
| width: 100%; |
| border-collapse: collapse; |
| font-size: 0.92rem; |
| margin: 0 0 1rem; |
| } |
| |
| th, |
| td { |
| border: 1px solid color-mix(in srgb, CanvasText 18%, transparent); |
| padding: 0.45rem 0.55rem; |
| text-align: left; |
| vertical-align: top; |
| } |
| |
| th { |
| font-weight: 600; |
| } |
| |
| .ecosystem { |
| font-size: 0.92rem; |
| margin: 0 0 1rem; |
| } |
| |
| .ecosystem li { |
| margin-bottom: 0.35rem; |
| } |
| |
| blockquote { |
| margin: 1rem 0 0; |
| padding-left: 0.85rem; |
| border-left: 3px solid color-mix(in srgb, CanvasText 25%, transparent); |
| font-style: italic; |
| opacity: 0.9; |
| } |
| </style> |
| </head> |
| <body> |
| <main> |
| <h1>Dali</h1> |
| <h3>The open verification layer for AI</h3> |
| <p class="lede"> |
| Dali is the open verification layer for AI: it creates, scores, and preserves |
| evidence so AI-assisted outputs can be independently verified, exchanged, and |
| replayed. Legal AI is the proving ground. |
| </p> |
|
|
| <hr /> |
|
|
| <h2>Open Artifacts</h2> |
| <div class="artifact"> |
| <strong |
| ><a href="https://huggingface.co/datasets/yenklabs/open-evidence-corpus" |
| >Dali Open Evidence Corpus</a |
| ></strong |
| > |
| <span class="status">· available</span> |
| <p>Cross-jurisdiction seed evidence for reproducible research.</p> |
| </div> |
| <div class="artifact"> |
| <strong |
| ><a href="https://huggingface.co/datasets/yenklabs/dali-citation-benchmark" |
| >Dali Citation Benchmark (Seed Sample)</a |
| ></strong |
| > |
| <span class="status">· available</span> |
| <p>5 hand-curated cases, 14 authorities — methodology review sample (not the full run).</p> |
| </div> |
| <div class="artifact"> |
| <strong |
| ><a href="https://huggingface.co/datasets/yenklabs/dali-verification-taxonomy" |
| >Dali Verification Taxonomy</a |
| ></strong |
| > |
| <span class="status">· available</span> |
| <p>Shared vocabulary for classifying verification outcomes.</p> |
| </div> |
| <div class="artifact"> |
| <strong |
| ><a href="https://github.com/yenklabs/Dali/tree/main/data/results" |
| >Full evaluation run</a |
| ></strong |
| > |
| <span class="status">· available</span> |
| <p>524 citations · 3 models · 5 jurisdiction tracks — on GitHub data/results/.</p> |
| </div> |
| <div class="artifact"> |
| <strong><a href="https://github.com/yenklabs/Dali">Evidence Infrastructure</a></strong> |
| <span class="status">· available (early-stage v0.2.1)</span> |
| <p>Open verification engine, methodologies, and reproducible evaluation workflows.</p> |
| </div> |
|
|
| <hr /> |
|
|
| <h2>Quick Start</h2> |
| <pre><code>from datasets import load_dataset |
|
|
| dataset = load_dataset("yenklabs/open-evidence-corpus")</code></pre> |
|
|
| <p class="cta"> |
| <a href="https://github.com/yenklabs/Dali">GitHub</a> · |
| <a href="https://github.com/yenklabs/Dali/blob/main/CONTRIBUTING.md">Contribute</a> · |
| <a href="https://yenklabs.com">yenklabs.com</a> |
| </p> |
|
|
| <hr /> |
|
|
| <h2>Research models (roadmap)</h2> |
| <p style="margin: 0 0 0.75rem; opacity: 0.85; font-size: 0.95rem"> |
| Lightweight, reproducible research models — not foundation models. All <em>planned</em>. |
| </p> |
| <table> |
| <thead> |
| <tr> |
| <th>Release</th> |
| <th>Model</th> |
| <th>Purpose</th> |
| <th>Status</th> |
| </tr> |
| </thead> |
| <tbody> |
| <tr> |
| <td>v0.1</td> |
| <td>Dali Verification Taxonomy Classifier</td> |
| <td>Predict standardized verification outcome labels</td> |
| <td>planned</td> |
| </tr> |
| <tr> |
| <td>v0.2</td> |
| <td>Dali Citation Risk Classifier</td> |
| <td>Estimate citation verification risk from evidence metadata</td> |
| <td>planned</td> |
| </tr> |
| <tr> |
| <td>v0.3</td> |
| <td>Dali Authority Matching Baseline</td> |
| <td>Reproducible baseline for authority matching experiments</td> |
| <td>planned</td> |
| </tr> |
| <tr> |
| <td>v0.4</td> |
| <td>Dali Proposition Support Classifier</td> |
| <td>Classify proposition support relationships for legal authorities</td> |
| <td>planned</td> |
| </tr> |
| </tbody> |
| </table> |
|
|
| <h2>Ecosystem</h2> |
| <ul class="ecosystem"> |
| <li><strong>Datasets</strong> — Open Evidence Corpus, Citation Benchmark seed sample, Verification Taxonomy · <em>available</em></li> |
| <li><strong>Full evaluation run</strong> — 524 citations on <a href="https://github.com/yenklabs/Dali/tree/main/data/results">github.com/yenklabs/Dali/data/results</a> · <em>available</em></li> |
| <li><strong>Models</strong> — Taxonomy Classifier, Citation Risk, Authority Matching, Proposition Support · <em>planned</em></li> |
| <li><strong>Spaces</strong> — Evidence Explorer, Benchmark Dashboard, Citation Verification Demo · <em>planned</em></li> |
| <li><strong>Future datasets</strong> — Evaluation Prompts, Replay Corpus, Evidence Artifacts · <em>planned</em></li> |
| </ul> |
| <blockquote> |
| Dali is the open verification layer for AI: it creates, scores, and preserves evidence so |
| AI-assisted outputs can be independently verified, exchanged, and replayed. |
| </blockquote> |
| </main> |
| </body> |
| </html> |
|
|
