index.html

<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>L2-Bench Pipeline</title>
<!-- The Space is served in an iframe on huggingface.co, which refuses to be framed
     within itself; without this every link below silently fails to navigate. -->
<base target="_blank">
<link rel="stylesheet" href="https://fonts.googleapis.com/css2?family=Open+Sans:wght@400;700&display=swap">
<style>
  :root { --oxford-blue: #011E41; --electric-blue: #0043E0; }
  body { margin: 0; background: #fff;
         font: 400 17px/1.8 'Open Sans', system-ui, sans-serif; color: var(--oxford-blue); }
  main { max-width: 34rem; margin: 0 auto; padding: 6rem 1.5rem; }
  p { margin: 0 0 2.5rem; }
  p:last-child { margin: 0; font-size: .9rem; }
  a { color: var(--electric-blue); }
  code { font-size: .9em; }
</style>
</head>
<body>
<main>
  <p>
    <strong>L2-Bench Pipeline</strong> is the reference implementation for
    <a href="https://huggingface.co/datasets/OUP/l2-bench">L2-Bench</a>, involving response
    generation and LLM-as-a-Judge scoring pipeline with <code>inspect-ai</code>.
  </p>

  <p>
    Full documentation for the pipeline, including prerequisites and all CLI options, is in
    the <a href="https://huggingface.co/spaces/OUP/l2-bench-pipeline/blob/main/README.md">README</a>.
  </p>

  <p>
    Code released under the
    <a href="https://huggingface.co/spaces/OUP/l2-bench-pipeline/blob/main/LICENSE">MIT Licence</a>.
    © 2026 Oxford University Press.
  </p>
</main>
</body>
</html>