ToolUse.1: leaderboard

README.md

@@ -1,10 +1,19 @@
 ---
 title: ToolUse.1
-emoji: 📊
-colorFrom: yellow
-colorTo: gray
+emoji: 🧰
+colorFrom: gray
+colorTo: red
 sdk: static
-pinned: false
+pinned: true
+license: apache-2.0
+short_description: Tool-calling leaderboard that scores abstention
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# ToolUse.1 — Leaderboard
+
+Static leaderboard for the [ToolUse.1](https://huggingface.co/datasets/HardcoreAI/ToolUse.1)
+tool-calling benchmark. It renders `results.json`, which is generated from harness result files
+by `publish.py` in the benchmark repository.
+
+Add a row by running the harness yourself and opening a discussion on the dataset repository
+with the raw results file.

app.js

@@ -0,0 +1,331 @@
+/* ==========================================================================
+   ToolUse.1 — shared client
+   Loads results (live results.json, else embedded snapshot), renders the
+   leaderboard, and drives the motion layer: reveal-on-scroll, count-up,
+   score bars, sortable columns.
+   ========================================================================== */
+
+"use strict";
+
+/* Gate motion behind a class so the no-JS view stays fully visible. */
+document.documentElement.classList.add("js");
+
+/* Inlined snapshot so pages render when opened straight from disk
+   (file:// blocks fetch). A served results.json overrides it. */
+const EMBEDDED_RESULTS = {
+  benchmark: "ToolUse.1",
+  version: "1.0.0",
+  updated: "2026-10-17",
+  corpus: { cases: 114, call_cases: 78, no_call_cases: 36, tools: 9, protocols: 3 },
+  rows: [
+    {
+      model: "oracle",
+      link: "",
+      protocol: "all",
+      score: 1.0,
+      score_ci95: [0.9889, 1.0],
+      correct_call: 1.0,
+      abstention: 1.0,
+      spurious_call_rate: 0.0,
+      n_cases: 342,
+      notes: "reads the expectations out of the corpus — grader reference",
+      self_check: true,
+      by_protocol: { json: 1.0, xml: 1.0, plain: 1.0 },
+    },
+    {
+      model: "HardcoreAI/Nova-1.1-0.8B",
+      link: "https://huggingface.co/HardcoreAI/Nova-1.1-0.8B",
+      protocol: "all",
+      score: 0.6404,
+      score_ci95: [0.5882, 0.6894],
+      correct_call: 0.5855,
+      abstention: 0.7593,
+      spurious_call_rate: 0.1667,
+      n_cases: 342,
+      notes: "greedy · thinking off · batch 8 · English-only · fine-tuned",
+      self_check: false,
+      by_protocol: { json: 0.6754, xml: 0.6316, plain: 0.614 },
+    },
+    {
+      model: "Qwen/Qwen3.5-0.8B",
+      link: "https://huggingface.co/Qwen/Qwen3.5-0.8B",
+      protocol: "all",
+      score: 0.5936,
+      score_ci95: [0.5408, 0.6443],
+      correct_call: 0.4829,
+      abstention: 0.8333,
+      spurious_call_rate: 0.0648,
+      n_cases: 342,
+      notes: "greedy · thinking off · batch 8 · English-only · base model",
+      self_check: false,
+      by_protocol: { json: 0.6754, xml: 0.6754, plain: 0.4298 },
+    },
+  ],
+};
+
+const pct = (v) => (typeof v === "number" ? (v * 100).toFixed(1) + "%" : "—");
+const ci = (lo, hi) => (lo == null || hi == null ? "—" : `${(lo * 100).toFixed(1)}% – ${(hi * 100).toFixed(1)}%`);
+
+const prefersReduced = window.matchMedia("(prefers-reduced-motion: reduce)").matches;
+
+/* ------------------------------------------------------------- motion */
+function revealOnScroll(root = document) {
+  const nodes = root.querySelectorAll(".reveal");
+  if (prefersReduced || !("IntersectionObserver" in window)) {
+    nodes.forEach((n) => n.classList.add("in"));
+    return;
+  }
+  const io = new IntersectionObserver(
+    (entries) => {
+      entries.forEach((entry) => {
+        if (entry.isIntersecting) {
+          entry.target.classList.add("in");
+          io.unobserve(entry.target);
+        }
+      });
+    },
+    { rootMargin: "0px 0px -8% 0px", threshold: 0.08 },
+  );
+  nodes.forEach((n) => io.observe(n));
+}
+
+function countUp(el) {
+  const target = parseFloat(el.dataset.count);
+  const suffix = el.dataset.suffix || "";
+  if (prefersReduced || Number.isNaN(target)) { el.textContent = `${target}${suffix}`; return; }
+  const start = performance.now();
+  const dur = 900;
+  const step = (now) => {
+    const t = Math.min(1, (now - start) / dur);
+    const eased = 1 - Math.pow(1 - t, 3);
+    el.textContent = `${Math.round(target * eased)}${suffix}`;
+    if (t < 1) requestAnimationFrame(step);
+    else el.textContent = `${target}${suffix}`;
+  };
+  requestAnimationFrame(step);
+}
+
+/* ------------------------------------------------------------- rendering */
+function scoreCell(value) {
+  const wrap = document.createElement("div");
+  wrap.className = "score-cell";
+  const bar = document.createElement("div");
+  bar.className = "bar";
+  const fill = document.createElement("i");
+  fill.dataset.w = Math.max(0, Math.min(1, value || 0));
+  bar.append(fill);
+  const val = document.createElement("span");
+  val.className = "score-value";
+  val.textContent = pct(value);
+  wrap.append(bar, val);
+  return wrap;
+}
+
+function spuriousCell(value) {
+  const cell = document.createElement("td");
+  cell.className = "num";
+  if (typeof value !== "number") { cell.textContent = "—"; return cell; }
+  cell.textContent = pct(value);
+  cell.classList.add(value >= 0.3 ? "spurious-high" : value >= 0.1 ? "spurious-mid" : "");
+  return cell;
+}
+
+function protoMini(byProtocol) {
+  const wrap = document.createElement("div");
+  wrap.className = "proto-mini";
+  ["json", "xml", "plain"].forEach((key) => {
+    const box = document.createElement("div");
+    const label = document.createElement("span");
+    label.className = "pl";
+    label.textContent = key;
+    const bar = document.createElement("div");
+    bar.className = "pb";
+    const fill = document.createElement("i");
+    fill.dataset.w = Math.max(0, Math.min(1, byProtocol?.[key] ?? 0));
+    bar.append(fill);
+    const val = document.createElement("span");
+    val.className = "pv";
+    val.textContent = pct(byProtocol?.[key]);
+    box.append(label, bar, val);
+    wrap.append(box);
+  });
+  return wrap;
+}
+
+function rowNode(row, index, opts = {}) {
+  const { withProtocol = false, withNotes = false } = opts;
+  const tr = document.createElement("tr");
+  tr.dataset.score = row.score ?? 0;
+  tr.dataset.correct_call = row.correct_call ?? 0;
+  tr.dataset.abstention = row.abstention ?? 0;
+  tr.dataset.spurious_call_rate = row.spurious_call_rate ?? 0;
+  tr.dataset.n_cases = row.n_cases ?? 0;
+
+  const rank = document.createElement("td");
+  rank.className = "col-rank";
+  const badge = document.createElement("span");
+  badge.className = `rank-badge${index < 3 ? ` rank-${index + 1}` : ""}`;
+  badge.textContent = index + 1;
+  rank.append(badge);
+  tr.append(rank);
+
+  const model = document.createElement("td");
+  model.className = "col-model";
+  if (row.link) {
+    const a = document.createElement("a");
+    a.className = "model-link model-name";
+    a.href = row.link;
+    a.target = "_blank";
+    a.rel = "noopener";
+    a.textContent = row.model;
+    model.append(a);
+  } else {
+    const s = document.createElement("span");
+    s.className = "model-name";
+    s.textContent = row.model;
+    model.append(s);
+  }
+  const sub = [row.protocol === "all" ? "all protocols" : `protocol: ${row.protocol}`, row.notes].filter(Boolean).join(" · ");
+  if (sub) {
+    const s = document.createElement("span");
+    s.className = "model-sub";
+    s.textContent = sub;
+    model.append(s);
+  }
+  tr.append(model);
+
+  const score = document.createElement("td");
+  score.className = "num";
+  score.append(scoreCell(row.score));
+  const c = document.createElement("span");
+  c.className = "ci";
+  c.textContent = Array.isArray(row.score_ci95) ? ci(row.score_ci95[0], row.score_ci95[1]) : "";
+  score.append(c);
+  tr.append(score);
+
+  tr.append(Object.assign(document.createElement("td"), { className: "num", textContent: pct(row.correct_call) }));
+  tr.append(Object.assign(document.createElement("td"), { className: "num", textContent: pct(row.abstention) }));
+  tr.append(spuriousCell(row.spurious_call_rate));
+  tr.append(Object.assign(document.createElement("td"), { className: "num", textContent: row.n_cases ?? "—" }));
+
+  if (withProtocol) {
+    const proto = document.createElement("td");
+    proto.append(protoMini(row.by_protocol));
+    tr.append(proto);
+  }
+
+  if (withNotes) {
+    const notes = document.createElement("td");
+    notes.className = "col-notes";
+    if (row.self_check) {
+      const tag = document.createElement("span");
+      tag.className = "tag";
+      tag.textContent = "grader self-check";
+      notes.append(tag);
+    } else if (row.notes) {
+      const span = document.createElement("span");
+      span.className = "ci";
+      span.textContent = row.notes;
+      notes.append(span);
+    }
+    tr.append(notes);
+  }
+
+  return tr;
+}
+
+function renderBoard(data, host) {
+  const tbody = host.querySelector("[data-board-body]");
+  const stamp = host.querySelector("[data-board-stamp]");
+  const emptyEl = host.querySelector("[data-board-empty]");
+  const rows = [...(data.rows || [])].sort(
+    (a, b) => (b.score ?? 0) - (a.score ?? 0) || (a.spurious_call_rate ?? 1) - (b.spurious_call_rate ?? 1),
+  );
+
+  const hasProtocolCol = !!host.querySelector("[data-col-protocol]");
+  const hasNotesCol = !!host.querySelector("[data-col-notes]");
+
+  tbody.replaceChildren(...rows.map((row, i) => rowNode(row, i, { withProtocol: hasProtocolCol, withNotes: hasNotesCol })));
+
+  if (stamp) {
+    stamp.textContent = [
+      `${rows.length} entr${rows.length === 1 ? "y" : "ies"}`,
+      data.version ? `v${data.version}` : "",
+      data.updated ? `updated ${data.updated}` : "",
+    ].filter(Boolean).join(" · ");
+  }
+  if (emptyEl) emptyEl.hidden = rows.length > 0;
+
+  requestAnimationFrame(() => {
+    tbody.querySelectorAll("tr").forEach((tr, i) => {
+      setTimeout(() => tr.classList.add("in"), prefersReduced ? 0 : 40 * i);
+    });
+    tbody.querySelectorAll(".bar > i, .pb > i").forEach((fill) => {
+      fill.style.width = `${Math.max(0, Math.min(1, parseFloat(fill.dataset.w) || 0)) * 100}%`;
+    });
+  });
+}
+
+/* ------------------------------------------------------------- sorting */
+function wireSorting(host) {
+  const headers = host.querySelectorAll("th[data-sort]");
+  headers.forEach((th) => {
+    th.classList.add("sortable");
+    if (!th.querySelector(".arrow")) th.insertAdjacentHTML("beforeend", ' <span class="arrow">↑↓</span>');
+    th.addEventListener("click", () => {
+      const key = th.dataset.sort;
+      const dir = th.getAttribute("aria-sort") === "descending" ? "ascending" : "descending";
+      headers.forEach((h) => h.removeAttribute("aria-sort"));
+      th.setAttribute("aria-sort", dir);
+      const tbody = host.querySelector("[data-board-body]");
+      const glyph = th.querySelector(".arrow");
+      if (glyph) glyph.textContent = dir === "descending" ? " ↓" : " ↑";
+      const rows = [...tbody.querySelectorAll("tr")];
+      rows.sort((a, b) => {
+        const av = parseFloat(a.dataset[key]) || 0;
+        const bv = parseFloat(b.dataset[key]) || 0;
+        return dir === "descending" ? bv - av : av - bv;
+      });
+      tbody.replaceChildren(...rows);
+    });
+  });
+}
+
+/* ------------------------------------------------------------- bootstrap */
+async function loadResults() {
+  try {
+    const res = await fetch("results.json", { cache: "no-store" });
+    if (!res.ok) throw new Error(`HTTP ${res.status}`);
+    return await res.json();
+  } catch {
+    return EMBEDDED_RESULTS;
+  }
+}
+
+function applyMetricCards(data) {
+  const best = [...(data.rows || [])].sort((a, b) => (b.score ?? 0) - (a.score ?? 0))[0];
+  if (!best) return;
+  const map = { overall: best.score, call: best.correct_call, abstain: best.abstention };
+  document.querySelectorAll("[data-metric]").forEach((el) => {
+    const v = map[el.dataset.metric];
+    if (typeof v === "number") el.textContent = pct(v);
+  });
+}
+
+document.addEventListener("DOMContentLoaded", async () => {
+  document.querySelectorAll("[data-count]").forEach(countUp);
+  revealOnScroll();
+
+  const boards = document.querySelectorAll("[data-board]");
+  if (!boards.length) return;
+
+  const data = await loadResults();
+  boards.forEach((host) => {
+    renderBoard(data, host);
+    wireSorting(host);
+  });
+  applyMetricCards(data);
+
+  const year = document.getElementById("year");
+  if (year) year.textContent = new Date().getFullYear();
+});

index.html

@@ -1,19 +1,171 @@
 <!doctype html>
-<html>
-	<head>
-		<meta charset="utf-8" />
-		<meta name="viewport" content="width=device-width" />
-		<title>My static Space</title>
-		<link rel="stylesheet" href="style.css" />
-	</head>
-	<body>
-		<div class="card">
-			<h1>Welcome to your static Space!</h1>
-			<p>You can modify this app directly by editing <i>index.html</i> in the Files and versions tab.</p>
-			<p>
-				Also don't forget to check the
-				<a href="https://huggingface.co/docs/hub/spaces" target="_blank">Spaces documentation</a>.
-			</p>
-		</div>
-	</body>
+<html lang="en">
+<head>
+<meta charset="utf-8" />
+<meta name="viewport" content="width=device-width, initial-scale=1" />
+<title>ToolUse.1 — Tool-Calling Benchmark with Abstention</title>
+<meta name="description" content="ToolUse.1 measures whether a model calls tools correctly — and only when it should. 114 English cases · JSON / XML / plain-text protocols · deterministic grading, no LLM judge." />
+<link rel="icon" href="data:image/svg+xml,<svg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 32 32'><rect width='32' height='32' rx='7' fill='%230a0b0d'/><path d='M8 22V10h3.4v9.2H21V22z' fill='%23e5484d'/></svg>" />
+<link rel="stylesheet" href="styles.css" />
+</head>
+<body>
+
+<nav class="nav">
+  <div class="wrap">
+    <a class="brand" href="index.html"><span class="dot"></span>ToolUse<span class="mark">.1</span></a>
+    <div class="nav-links">
+      <a href="index.html" aria-current="page">Overview</a>
+      <a href="leaderboard.html">Leaderboard</a>
+      <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1" target="_blank" rel="noopener">Dataset</a>
+      <a class="nav-cta" href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/discussions" target="_blank" rel="noopener">Submit a model</a>
+    </div>
+  </div>
+</nav>
+
+<header class="hero">
+  <div class="wrap">
+    <p class="eyebrow anim" style="animation-delay:.05s">Hardcore AI · benchmark</p>
+    <h1 class="anim" style="animation-delay:.12s">ToolUse<span class="mark">.1</span></h1>
+    <p class="lede anim" style="animation-delay:.2s">
+      Can a model call a tool when it should — and stay quiet when it shouldn't?
+      ToolUse.1 scores correct invocation, correct abstention, and correct clarification
+      separately, on 114 authored cases in English.
+    </p>
+    <ul class="stats anim" style="animation-delay:.3s">
+      <li><b data-count="114">114</b><span>cases</span></li>
+      <li><b data-count="78">78</b><span>require a call</span></li>
+      <li><b data-count="36">36</b><span>abstain / clarify</span></li>
+      <li><b data-count="9">9</b><span>tool schemas</span></li>
+      <li><b data-count="3">3</b><span>protocols</span></li>
+      <li><b data-count="0">0</b><span>LLM judges</span></li>
+    </ul>
+  </div>
+</header>
+
+<main class="wrap">
+
+  <section class="metrics reveal">
+    <div class="metric">
+      <div class="label">Invocation</div>
+      <div class="value" data-metric="overall">—</div>
+      <div class="note">Right tool, semantically correct arguments. Expressions are evaluated — <code>45+12</code> equals <code>45 + 12</code>.</div>
+    </div>
+    <div class="metric">
+      <div class="label">Abstention</div>
+      <div class="value" data-metric="abstain">—</div>
+      <div class="note">No call when no tool applies: knowledge questions, absent capabilities, explicit no-tool instructions.</div>
+    </div>
+    <div class="metric">
+      <div class="label">Clarification</div>
+      <div class="value" data-metric="call">—</div>
+      <div class="note">A required argument is missing. Asking the user passes; guessing an argument fails.</div>
+    </div>
+  </section>
+
+  <section class="grid reveal">
+    <div class="card">
+      <div class="k">What it measures</div>
+      <h3>Right call, right time</h3>
+      <p>Nine tools across arithmetic, weather, search, translation, conversion, timers, notes and time. Every case is a realistic user turn — no trivia, no synthetic templates.</p>
+    </div>
+    <div class="card">
+      <div class="k">Three protocols</div>
+      <h3>JSON · XML · plain-text</h3>
+      <p>The output protocol is a run-time axis, not a data axis. Each case is rendered in all three; grading uses canonical call semantics, so a model is judged on intent, not formatting.</p>
+    </div>
+    <div class="card">
+      <div class="k">Deterministic grading</div>
+      <h3>No LLM judge</h3>
+      <p>Arguments are compared as values with alias tables for cities, currencies, units and timezones. Exactly one call is required for call cases. The oracle self-check scores 100%.</p>
+    </div>
+  </section>
+
+  <section class="reveal">
+    <h2 class="section">Leaderboard — top of the board</h2>
+    <div class="board" data-board>
+      <div class="board-head">
+        <h2>ToolUse.1 — Leaderboard</h2>
+        <p class="stamp" data-board-stamp>loading results…</p>
+      </div>
+      <table>
+        <thead>
+          <tr>
+            <th class="col-rank">#</th>
+            <th class="col-model">Model</th>
+            <th class="num" title="Passed cases / all cases. Wilson 95% interval shown underneath.">Score</th>
+            <th class="num" title="Correct call when a tool applies">Correct call</th>
+            <th class="num" title="Correct abstention when no tool applies">Abstention</th>
+            <th class="num" title="No-call cases where any call was emitted. Lower is better.">Spurious ↓</th>
+            <th class="num" title="Cases scored">n</th>
+          </tr>
+        </thead>
+        <tbody data-board-body>
+          <tr class="in" data-score="1" data-correct_call="1" data-abstention="1" data-spurious_call_rate="0" data-n_cases="342">
+            <td class="col-rank"><span class="rank-badge rank-1">1</span></td>
+            <td class="col-model"><span class="model-name">oracle</span><span class="model-sub">all protocols · grader reference</span></td>
+            <td class="num"><div class="score-cell"><div class="bar"><i style="width:100%"></i></div><span class="score-value">100.0%</span></div><span class="ci">98.9% – 100.0%</span></td>
+            <td class="num">100.0%</td>
+            <td class="num">100.0%</td>
+            <td class="num">0.0%</td>
+            <td class="num">342</td>
+          </tr>
+          <tr class="in" data-score="0.6404" data-correct_call="0.5855" data-abstention="0.7593" data-spurious_call_rate="0.1667" data-n_cases="342">
+            <td class="col-rank"><span class="rank-badge rank-2">2</span></td>
+            <td class="col-model"><a class="model-link model-name" href="https://huggingface.co/HardcoreAI/Nova-1.1-0.8B" target="_blank" rel="noopener">HardcoreAI/Nova-1.1-0.8B</a><span class="model-sub">all protocols · greedy · thinking off · fine-tuned</span></td>
+            <td class="num"><div class="score-cell"><div class="bar"><i style="width:64.0%"></i></div><span class="score-value">64.0%</span></div><span class="ci">58.8% – 68.9%</span></td>
+            <td class="num">58.6%</td>
+            <td class="num">75.9%</td>
+            <td class="num spurious-mid">16.7%</td>
+            <td class="num">342</td>
+          </tr>
+          <tr class="in" data-score="0.5936" data-correct_call="0.4829" data-abstention="0.8333" data-spurious_call_rate="0.0648" data-n_cases="342">
+            <td class="col-rank"><span class="rank-badge rank-3">3</span></td>
+            <td class="col-model"><a class="model-link model-name" href="https://huggingface.co/Qwen/Qwen3.5-0.8B" target="_blank" rel="noopener">Qwen/Qwen3.5-0.8B</a><span class="model-sub">all protocols · greedy · thinking off · base model</span></td>
+            <td class="num"><div class="score-cell"><div class="bar"><i style="width:59.4%"></i></div><span class="score-value">59.4%</span></div><span class="ci">54.1% – 64.4%</span></td>
+            <td class="num">48.3%</td>
+            <td class="num">83.3%</td>
+            <td class="num">6.5%</td>
+            <td class="num">342</td>
+          </tr>
+        </tbody>
+      </table>
+      <p class="empty" data-board-empty hidden>No model results yet. See <strong>Submit a model</strong> below.</p>
+    </div>
+    <p style="margin:1rem 0 0"><a href="leaderboard.html">Open the full leaderboard →</a></p>
+  </section>
+
+  <section class="submit reveal">
+    <h2 class="section">Submit a model</h2>
+    <p>Run the benchmark yourself, keep the raw results JSON the harness writes, and open a discussion on the dataset repository with the model id and configuration.</p>
+    <pre><code>git clone https://huggingface.co/datasets/HardcoreAI/ToolUse.1
+pip install -r requirements.txt         # transformers 5.x for Nova / Qwen3.5
+
+python -m tooluse1 selftest             # oracle must score exactly 100%
+python -m tooluse1 run --model &lt;id&gt; \
+    --backend transformers --protocol all --batch-size 8 --out results/</code></pre>
+    <ol class="steps">
+      <li><b>Verify the harness</b> — run the self-check first; it validates the grader against the corpus.</li>
+      <li><b>Run all three protocols</b> and keep the results JSON the harness emits.</li>
+      <li><b>Open a discussion</b> on <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/discussions" target="_blank" rel="noopener">HardcoreAI/ToolUse.1</a> with the model id, protocols, case count, decoding config, and the raw JSON attachment.</li>
+    </ol>
+    <p style="margin-top:1rem;color:var(--dim);font-size:0.86rem;">
+      Recommended: greedy decoding (<code>do_sample=False</code>), <code>max_new_tokens=256</code>, unmodified system prompt,
+      no few-shot examples, reasoning mode disabled where supported. One dtype per row — bfloat16 on Ampere+, float16 on T4.
+    </p>
+  </section>
+
+</main>
+
+<footer>
+  <div class="wrap">
+    <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1">Dataset</a>
+    <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/blob/main/SPEC.md">Specification</a>
+    <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/discussions">Discussions</a>
+    <a href="https://huggingface.co/HardcoreAI">Hardcore AI</a>
+    <span class="license">Apache-2.0 · v1.0.0 · <span id="year">2026</span></span>
+  </div>
+</footer>
+
+<script src="app.js"></script>
+</body>
 </html>

leaderboard.html

@@ -0,0 +1,156 @@
+<!doctype html>
+<html lang="en">
+<head>
+<meta charset="utf-8" />
+<meta name="viewport" content="width=device-width, initial-scale=1" />
+<title>Leaderboard — ToolUse.1</title>
+<meta name="description" content="Model leaderboard for ToolUse.1: overall score, correct invocation, correct abstention and spurious-call rate across JSON, XML and plain-text protocols." />
+<link rel="icon" href="data:image/svg+xml,<svg xmlns='http://www.w3.org/2000/svg' viewBox='0 0 32 32'><rect width='32' height='32' rx='7' fill='%230a0b0d'/><path d='M8 22V10h3.4v9.2H21V22z' fill='%23e5484d'/></svg>" />
+<link rel="stylesheet" href="styles.css" />
+</head>
+<body>
+
+<nav class="nav">
+  <div class="wrap">
+    <a class="brand" href="index.html"><span class="dot"></span>ToolUse<span class="mark">.1</span></a>
+    <div class="nav-links">
+      <a href="index.html">Overview</a>
+      <a href="leaderboard.html" aria-current="page">Leaderboard</a>
+      <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1" target="_blank" rel="noopener">Dataset</a>
+      <a class="nav-cta" href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/discussions" target="_blank" rel="noopener">Submit a model</a>
+    </div>
+  </div>
+</nav>
+
+<header class="hero" style="padding:3.4rem 0 1.2rem">
+  <div class="wrap">
+    <p class="eyebrow anim" style="animation-delay:.05s">ToolUse.1 · results</p>
+    <h1 class="anim" style="animation-delay:.12s">Leader<span class="mark">board</span></h1>
+    <p class="lede anim" style="animation-delay:.2s">
+      One row per submitted run. Sorted by overall score; click any numeric header to re-rank.
+      Every entry reports invocation, abstention and calibration — a high score with a high
+      spurious-call rate is not the same model as a high score with a clean one.
+    </p>
+  </div>
+</header>
+
+<main class="wrap" style="padding-top:1.6rem">
+
+  <section class="reveal">
+    <div class="board" data-board>
+      <div class="board-head">
+        <h2>All submissions</h2>
+        <div class="board-tools">
+          <p class="stamp" data-board-stamp>loading results…</p>
+        </div>
+      </div>
+      <table>
+        <thead>
+          <tr>
+            <th class="col-rank">#</th>
+            <th class="col-model">Model</th>
+            <th class="num" data-sort="score" title="Passed cases / all cases. Wilson 95% interval underneath.">Score <span class="arrow">↑↓</span></th>
+            <th class="num" data-sort="correct_call" title="Correct call when a tool applies">Correct call <span class="arrow">↑↓</span></th>
+            <th class="num" data-sort="abstention" title="Correct abstention when no tool applies">Abstention <span class="arrow">↑↓</span></th>
+            <th class="num" data-sort="spurious_call_rate" title="No-call cases where any call was emitted. Lower is better.">Spurious ↓ <span class="arrow">↑↓</span></th>
+            <th class="num" data-sort="n_cases" title="Cases scored">n <span class="arrow">↑↓</span></th>
+            <th data-col-protocol>By protocol</th>
+            <th class="col-notes" data-col-notes>Notes</th>
+          </tr>
+        </thead>
+        <tbody data-board-body>
+          <tr class="in" data-score="1" data-correct_call="1" data-abstention="1" data-spurious_call_rate="0" data-n_cases="342">
+            <td class="col-rank"><span class="rank-badge rank-1">1</span></td>
+            <td class="col-model"><span class="model-name">oracle</span><span class="model-sub">all protocols · reads the expectations out of the corpus</span></td>
+            <td class="num"><div class="score-cell"><div class="bar"><i style="width:100%"></i></div><span class="score-value">100.0%</span></div><span class="ci">98.9% – 100.0%</span></td>
+            <td class="num">100.0%</td>
+            <td class="num">100.0%</td>
+            <td class="num">0.0%</td>
+            <td class="num">342</td>
+            <td><div class="proto-mini">
+              <div><span class="pl">json</span><div class="pb"><i style="width:100%"></i></div><span class="pv">100.0%</span></div>
+              <div><span class="pl">xml</span><div class="pb"><i style="width:100%"></i></div><span class="pv">100.0%</span></div>
+              <div><span class="pl">plain</span><div class="pb"><i style="width:100%"></i></div><span class="pv">100.0%</span></div>
+            </div></td>
+            <td class="col-notes"><span class="tag">grader self-check</span></td>
+          </tr>
+          <tr class="in" data-score="0.6404" data-correct_call="0.5855" data-abstention="0.7593" data-spurious_call_rate="0.1667" data-n_cases="342">
+            <td class="col-rank"><span class="rank-badge rank-2">2</span></td>
+            <td class="col-model"><a class="model-link model-name" href="https://huggingface.co/HardcoreAI/Nova-1.1-0.8B" target="_blank" rel="noopener">HardcoreAI/Nova-1.1-0.8B</a><span class="model-sub">all protocols · greedy · thinking off · fine-tuned</span></td>
+            <td class="num"><div class="score-cell"><div class="bar"><i style="width:64.0%"></i></div><span class="score-value">64.0%</span></div><span class="ci">58.8% – 68.9%</span></td>
+            <td class="num">58.6%</td>
+            <td class="num">75.9%</td>
+            <td class="num spurious-mid">16.7%</td>
+            <td class="num">342</td>
+            <td><div class="proto-mini">
+              <div><span class="pl">json</span><div class="pb"><i style="width:67.5%"></i></div><span class="pv">67.5%</span></div>
+              <div><span class="pl">xml</span><div class="pb"><i style="width:63.2%"></i></div><span class="pv">63.2%</span></div>
+              <div><span class="pl">plain</span><div class="pb"><i style="width:61.4%"></i></div><span class="pv">61.4%</span></div>
+            </div></td>
+            <td class="col-notes"><span class="ci">fine-tuned · bf16 · batch 8</span></td>
+          </tr>
+          <tr class="in" data-score="0.5936" data-correct_call="0.4829" data-abstention="0.8333" data-spurious_call_rate="0.0648" data-n_cases="342">
+            <td class="col-rank"><span class="rank-badge rank-3">3</span></td>
+            <td class="col-model"><a class="model-link model-name" href="https://huggingface.co/Qwen/Qwen3.5-0.8B" target="_blank" rel="noopener">Qwen/Qwen3.5-0.8B</a><span class="model-sub">all protocols · greedy · thinking off · base model</span></td>
+            <td class="num"><div class="score-cell"><div class="bar"><i style="width:59.4%"></i></div><span class="score-value">59.4%</span></div><span class="ci">54.1% – 64.4%</span></td>
+            <td class="num">48.3%</td>
+            <td class="num">83.3%</td>
+            <td class="num">6.5%</td>
+            <td class="num">342</td>
+            <td><div class="proto-mini">
+              <div><span class="pl">json</span><div class="pb"><i style="width:67.5%"></i></div><span class="pv">67.5%</span></div>
+              <div><span class="pl">xml</span><div class="pb"><i style="width:67.5%"></i></div><span class="pv">67.5%</span></div>
+              <div><span class="pl">plain</span><div class="pb"><i style="width:43.0%"></i></div><span class="pv">43.0%</span></div>
+            </div></td>
+            <td class="col-notes"><span class="ci">base model · bf16 · batch 8</span></td>
+          </tr>
+        </tbody>
+      </table>
+      <p class="empty" data-board-empty hidden>No model results yet. Open a discussion to submit the first one.</p>
+    </div>
+  </section>
+
+  <section class="grid reveal">
+    <div class="card">
+      <div class="k">Score</div>
+      <h3>What counts</h3>
+      <p>A case passes when the emitted call set matches the expectation: the right tool, semantically correct arguments, and — on no-call cases — nothing emitted at all. Wilson 95% interval in parentheses.</p>
+    </div>
+    <div class="card">
+      <div class="k">Correct call ↑</div>
+      <h3>Recall on tool cases</h3>
+      <p>Share of call-eligible cases where the model produced the expected invocation with correct arguments. High abstention at low correct-call means the model is too shy.</p>
+    </div>
+    <div class="card">
+      <div class="k">Spurious ↓</div>
+      <h3>Calibration</h3>
+      <p>Share of no-call cases where the model emitted *some* call anyway. High abstention at high spurious means it is noisy, not cautious — the two are not interchangeable.</p>
+    </div>
+  </section>
+
+  <section class="submit reveal">
+    <h2 class="section">Submit a model</h2>
+    <p>Run the harness, keep the raw results JSON, and open a discussion on the dataset repository. Rows are added by hand — include the model id, protocols, case count, decoding configuration and the JSON attachment.</p>
+    <pre><code>python -m tooluse1 selftest    # oracle must score exactly 100%
+python -m tooluse1 run --model &lt;id&gt; --backend transformers \
+    --protocol all --batch-size 8 --out results/</code></pre>
+    <p style="margin:1rem 0 0">
+      <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/discussions" target="_blank" rel="noopener">Open a submission discussion →</a>
+    </p>
+  </section>
+
+</main>
+
+<footer>
+  <div class="wrap">
+    <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1">Dataset</a>
+    <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/blob/main/SPEC.md">Specification</a>
+    <a href="https://huggingface.co/datasets/HardcoreAI/ToolUse.1/discussions">Discussions</a>
+    <a href="https://huggingface.co/HardcoreAI">Hardcore AI</a>
+    <span class="license">Apache-2.0 · v1.0.0 · <span id="year">2026</span></span>
+  </div>
+</footer>
+
+<script src="app.js"></script>
+</body>
+</html>

results.json

@@ -0,0 +1,105 @@
+{
+  "benchmark": "ToolUse.1",
+  "version": "1.0.0",
+  "updated": "2026-10-17",
+  "corpus": {
+    "cases": 114,
+    "call_cases": 78,
+    "no_call_cases": 36,
+    "by_language": {
+      "en": 114
+    },
+    "by_category": {
+      "calculator": 15,
+      "get_weather": 8,
+      "web_search": 8,
+      "translate": 8,
+      "convert_currency": 8,
+      "convert_units": 8,
+      "set_timer": 6,
+      "save_note": 6,
+      "get_time": 6,
+      "negative": 30,
+      "clarification": 5,
+      "multi_step": 6
+    },
+    "sha256": "17a6ef84c22648d83add6f56a03b24f896420837f4e69039c22d61cf8ad2c3e1"
+  },
+  "rows": [
+    {
+      "model": "oracle",
+      "link": "",
+      "protocol": "all",
+      "score": 1.0,
+      "score_ci95": [
+        0.9889,
+        1.0
+      ],
+      "correct_call": 1.0,
+      "abstention": 1.0,
+      "spurious_call_rate": 0.0,
+      "n_cases": 342,
+      "notes": "reads the expectations out of the corpus — grader reference",
+      "self_check": true,
+      "by_protocol": {
+        "json": 1.0,
+        "plain": 1.0,
+        "xml": 1.0
+      },
+      "config": {
+        "backend": "harness self-check",
+        "results_file": "selftest__oracle__all.json"
+      }
+    },
+    {
+      "model": "HardcoreAI/Nova-1.1-0.8B",
+      "link": "",
+      "protocol": "all",
+      "score": 0.6404,
+      "score_ci95": [
+        0.5882,
+        0.6894
+      ],
+      "correct_call": 0.5855,
+      "abstention": 0.7593,
+      "spurious_call_rate": 0.1667,
+      "n_cases": 342,
+      "notes": "Nova-1.1-0.8B | greedy | thinking off | batch 8 | transformers 5.5.4 | English-only ToolUse.1",
+      "self_check": false,
+      "by_protocol": {
+        "json": 0.6754,
+        "plain": 0.614,
+        "xml": 0.6316
+      },
+      "config": {
+        "backend": "external",
+        "results_file": "HardcoreAI__Nova-1.1-0.8B__all.json"
+      }
+    },
+    {
+      "model": "Qwen/Qwen3.5-0.8B",
+      "link": "",
+      "protocol": "all",
+      "score": 0.5936,
+      "score_ci95": [
+        0.5408,
+        0.6443
+      ],
+      "correct_call": 0.4829,
+      "abstention": 0.8333,
+      "spurious_call_rate": 0.0648,
+      "n_cases": 342,
+      "notes": "Qwen/Qwen3.5-0.8B | greedy | thinking off | batch 8 | transformers 5.5.4 | English-only ToolUse.1 | base model",
+      "self_check": false,
+      "by_protocol": {
+        "json": 0.6754,
+        "plain": 0.4298,
+        "xml": 0.6754
+      },
+      "config": {
+        "backend": "external",
+        "results_file": "Qwen__Qwen3.5-0.8B__all.json"
+      }
+    }
+  ]
+}

styles.css

@@ -0,0 +1,298 @@
+/* ==========================================================================
+   ToolUse.1 — design system
+   Dark, editorial, one accent. Motion is intentional, not decorative:
+   entrance cascade, count-up, bars that grow, sections that reveal.
+   ========================================================================== */
+
+:root {
+  --ink: #06070a;
+  --ink-2: #0a0c10;
+  --panel: #111419;
+  --panel-2: #14171d;
+  --line: #1f242e;
+  --line-soft: #171b22;
+  --text: #e9ebf0;
+  --muted: #8d95a6;
+  --dim: #626b7d;
+  --accent: #e5484d;
+  --accent-soft: #ff7a70;
+  --accent-glow: rgba(229, 72, 77, 0.35);
+  --good: #4ade80;
+  --warn: #ffb454;
+  --bronze: #c98a5b;
+  --mono: ui-monospace, "SF Mono", "JetBrains Mono", Menlo, Consolas, monospace;
+  --sans: system-ui, -apple-system, "Segoe UI", Roboto, "Helvetica Neue", sans-serif;
+  --serif: "Iowan Old Style", "Palatino Linotype", Palatino, Georgia, ui-serif, serif;
+  --ease: cubic-bezier(0.22, 1, 0.36, 1);
+  --radius: 12px;
+  --shadow: 0 18px 50px -28px rgba(0, 0, 0, 0.9);
+}
+
+* { box-sizing: border-box; }
+html { -webkit-text-size-adjust: 100%; scroll-behavior: smooth; }
+
+body {
+  margin: 0;
+  min-height: 100vh;
+  background-color: var(--ink);
+  background-image:
+    radial-gradient(1200px 560px at 10% -14%, rgba(229, 72, 77, 0.16), transparent 60%),
+    radial-gradient(980px 500px at 98% 2%, rgba(96, 118, 214, 0.11), transparent 56%),
+    radial-gradient(760px 640px at 50% 120%, rgba(229, 72, 77, 0.08), transparent 60%);
+  background-attachment: fixed;
+  color: var(--text);
+  font-family: var(--sans);
+  font-size: 16px;
+  line-height: 1.55;
+  -webkit-font-smoothing: antialiased;
+  text-rendering: optimizeLegibility;
+  overflow-x: hidden;
+}
+
+/* fine grain overlay — keeps the large dark fields from banding */
+body::after {
+  content: "";
+  position: fixed;
+  inset: 0;
+  pointer-events: none;
+  z-index: 1;
+  opacity: 0.035;
+  background-image: url("data:image/svg+xml,%3Csvg xmlns='http://www.w3.org/2000/svg' width='160' height='160'%3E%3Cfilter id='n'%3E%3CfeTurbulence type='fractalNoise' baseFrequency='0.8' numOctaves='3'/%3E%3C/filter%3E%3Crect width='100%25' height='100%25' filter='url(%23n)'/%3E%3C/svg%3E");
+}
+
+a { color: var(--text); border-bottom: 1px solid var(--line); text-decoration: none; transition: border-color 0.25s var(--ease), color 0.25s var(--ease); }
+a:hover { border-bottom-color: var(--accent); color: #fff; }
+
+.wrap { width: min(1120px, 100% - 2.5rem); margin-inline: auto; position: relative; z-index: 2; }
+
+/* ------------------------------------------------------------------ nav */
+.nav {
+  position: sticky;
+  top: 0;
+  z-index: 20;
+  backdrop-filter: blur(14px);
+  -webkit-backdrop-filter: blur(14px);
+  background: rgba(6, 7, 10, 0.72);
+  border-bottom: 1px solid var(--line-soft);
+}
+.nav .wrap { display: flex; align-items: center; gap: 1.4rem; height: 58px; }
+.brand {
+  display: inline-flex; align-items: center; gap: 0.55rem;
+  font-family: var(--mono); font-weight: 600; font-size: 0.92rem; letter-spacing: 0.02em;
+  border: none;
+}
+.brand .dot {
+  width: 9px; height: 9px; border-radius: 3px; background: var(--accent);
+  box-shadow: 0 0 0 3px rgba(229, 72, 77, 0.15);
+  animation: pulse 2.6s var(--ease) infinite;
+}
+@keyframes pulse {
+  0%, 100% { box-shadow: 0 0 0 3px rgba(229, 72, 77, 0.15); }
+  50% { box-shadow: 0 0 0 7px rgba(229, 72, 77, 0.04); }
+}
+.brand span.mark { color: var(--accent); }
+.nav-links { display: flex; gap: 1.25rem; margin-left: auto; align-items: center; }
+.nav-links a {
+  border: none; font-size: 0.84rem; color: var(--muted); letter-spacing: 0.01em;
+  position: relative; padding: 0.2rem 0;
+}
+.nav-links a::after {
+  content: ""; position: absolute; left: 0; right: 100%; bottom: -2px; height: 1px;
+  background: var(--accent); transition: right 0.3s var(--ease);
+}
+.nav-links a:hover { color: var(--text); }
+.nav-links a:hover::after, .nav-links a[aria-current="page"]::after { right: 0; }
+.nav-links a[aria-current="page"] { color: var(--text); }
+.nav-cta {
+  border: 1px solid var(--line) !important; border-radius: 999px; padding: 0.32rem 0.85rem !important;
+  color: var(--text) !important;
+}
+.nav-cta:hover { border-color: var(--accent) !important; }
+
+/* ------------------------------------------------------------- headings */
+.eyebrow {
+  margin: 0 0 0.9rem; font-family: var(--mono); font-size: 0.7rem;
+  letter-spacing: 0.22em; text-transform: uppercase; color: var(--dim);
+}
+h1 {
+  margin: 0 0 0.7rem; font-family: var(--serif); font-weight: 400;
+  font-size: clamp(2.8rem, 9vw, 5rem); line-height: 0.92; letter-spacing: -0.03em;
+}
+h1 .mark { color: var(--accent); }
+.lede { max-width: 58ch; margin: 0 0 1.6rem; font-size: 1.1rem; color: var(--muted); }
+
+h2.section {
+  margin: 0 0 1.1rem; font-size: 0.72rem; font-family: var(--mono);
+  letter-spacing: 0.2em; text-transform: uppercase; color: var(--muted); font-weight: 500;
+}
+
+/* ---------------------------------------------------------------- hero */
+header.hero { padding: 4.2rem 0 1.4rem; border-bottom: 1px solid var(--line); position: relative; }
+
+.stats { display: flex; flex-wrap: wrap; gap: 1.3rem 2.4rem; margin: 2rem 0 0.6rem; padding: 0; list-style: none; }
+.stats li { display: flex; flex-direction: column; gap: 0.14rem; }
+.stats b {
+  font-family: var(--mono); font-size: 1.55rem; font-weight: 500;
+  font-variant-numeric: tabular-nums; letter-spacing: -0.01em;
+  background: linear-gradient(180deg, #fff, #aab2c2);
+  -webkit-background-clip: text; background-clip: text; color: transparent;
+}
+.stats span { font-size: 0.64rem; letter-spacing: 0.16em; text-transform: uppercase; color: var(--dim); }
+
+/* ---------------------------------------------------------------- layout */
+main { padding: 2.8rem 0 4rem; display: grid; gap: 3.4rem; }
+.grid { display: grid; grid-template-columns: repeat(auto-fit, minmax(250px, 1fr)); gap: 1rem; }
+.split { display: grid; grid-template-columns: 1.35fr 1fr; gap: 1.2rem; align-items: start; }
+
+.card {
+  padding: 1.4rem 1.5rem; border: 1px solid var(--line); border-radius: var(--radius);
+  background: linear-gradient(180deg, var(--panel-2), var(--panel));
+  transition: transform 0.35s var(--ease), border-color 0.35s var(--ease), box-shadow 0.35s var(--ease);
+}
+.card:hover { transform: translateY(-3px); border-color: #2b3240; box-shadow: var(--shadow); }
+.card h3 { margin: 0 0 0.5rem; font-family: var(--serif); font-size: 1.2rem; font-weight: 400; }
+.card p { margin: 0; color: var(--muted); font-size: 0.92rem; }
+.card .k { font-family: var(--mono); font-size: 0.7rem; color: var(--accent-soft); letter-spacing: 0.1em; text-transform: uppercase; }
+
+/* ---------------------------------------------------------- leaderboard */
+.board { border: 1px solid var(--line); border-radius: var(--radius); background: var(--panel); overflow: hidden; box-shadow: var(--shadow); }
+.board-head {
+  padding: 1rem 1.25rem; border-bottom: 1px solid var(--line);
+  display: flex; align-items: baseline; justify-content: space-between; gap: 1rem; flex-wrap: wrap;
+}
+.board-head h2 { margin: 0; font-size: 0.92rem; font-family: var(--mono); letter-spacing: 0.06em; }
+.stamp { margin: 0; font-family: var(--mono); font-size: 0.7rem; color: var(--dim); }
+.board-tools { display: flex; gap: 0.4rem; align-items: center; }
+.chip {
+  font-family: var(--mono); font-size: 0.66rem; letter-spacing: 0.08em; text-transform: uppercase;
+  color: var(--muted); border: 1px solid var(--line); border-radius: 999px; padding: 0.24rem 0.7rem;
+  background: transparent; cursor: pointer; transition: all 0.25s var(--ease);
+}
+.chip:hover { color: var(--text); border-color: #2b3240; }
+.chip[aria-pressed="true"] { color: #fff; border-color: var(--accent); background: rgba(229, 72, 77, 0.12); }
+
+table { width: 100%; border-collapse: collapse; font-size: 0.9rem; }
+th {
+  padding: 0.7rem 0.9rem; text-align: left; font-family: var(--mono); font-size: 0.64rem;
+  font-weight: 500; letter-spacing: 0.14em; text-transform: uppercase; color: var(--dim);
+  border-bottom: 1px solid var(--line); white-space: nowrap; user-select: none;
+}
+th.sortable { cursor: pointer; transition: color 0.2s var(--ease); }
+th.sortable:hover { color: var(--text); }
+th .arrow { opacity: 0.35; margin-left: 0.25rem; }
+th[aria-sort] .arrow { opacity: 1; color: var(--accent); }
+td { padding: 0.8rem 0.9rem; border-bottom: 1px solid var(--line-soft); vertical-align: middle; }
+tr:last-child td { border-bottom: none; }
+tbody tr { transition: background 0.25s var(--ease); }
+.js tbody tr { opacity: 0; transform: translateY(8px); }
+.js tbody tr.in { opacity: 1; transform: none; }
+tbody tr:hover { background: var(--panel-2); }
+
+.num { text-align: right; font-family: var(--mono); font-variant-numeric: tabular-nums; white-space: nowrap; }
+.col-rank { width: 2.6rem; color: var(--dim); font-family: var(--mono); }
+.rank-badge {
+  display: inline-flex; align-items: center; justify-content: center; width: 1.55rem; height: 1.55rem;
+  border-radius: 999px; font-size: 0.72rem; border: 1px solid var(--line); color: var(--muted);
+}
+.rank-1 { border-color: #d8b25a; color: #f3d489; background: rgba(216, 178, 90, 0.1); box-shadow: 0 0 16px -4px rgba(216, 178, 90, 0.5); }
+.rank-2 { border-color: #9aa3b2; color: #c8ced8; background: rgba(154, 163, 178, 0.1); }
+.rank-3 { border-color: var(--bronze); color: #e0ad83; background: rgba(201, 138, 91, 0.1); }
+.col-model { min-width: 15rem; }
+.model-name { font-weight: 500; }
+.model-sub { display: block; margin-top: 0.16rem; font-family: var(--mono); font-size: 0.67rem; color: var(--dim); }
+
+.score-cell { display: flex; align-items: center; gap: 0.55rem; justify-content: flex-end; }
+.bar { position: relative; width: 96px; height: 5px; border-radius: 3px; background: #1b1f27; overflow: hidden; flex: none; }
+.bar > i {
+  position: absolute; inset: 0 auto 0 0; width: 0;
+  background: linear-gradient(90deg, var(--accent), var(--accent-soft));
+  border-radius: 3px; box-shadow: 0 0 14px -2px var(--accent-glow);
+  transition: width 1.1s var(--ease);
+}
+.score-value { font-family: var(--mono); font-variant-numeric: tabular-nums; font-weight: 600; min-width: 3.6rem; text-align: right; }
+.ci { font-family: var(--mono); font-size: 0.66rem; color: var(--dim); }
+.spurious-high { color: var(--accent); }
+.spurious-mid { color: var(--warn); }
+
+.tag {
+  display: inline-block; margin-right: 0.35rem; padding: 0.14rem 0.42rem; border: 1px solid var(--line);
+  border-radius: 999px; font-family: var(--mono); font-size: 0.58rem; letter-spacing: 0.1em;
+  text-transform: uppercase; color: var(--warn); white-space: nowrap;
+}
+.empty { margin: 0; padding: 1.2rem 1.25rem; color: var(--dim); font-style: italic; }
+.col-notes { max-width: 22rem; font-size: 0.76rem; color: var(--muted); }
+
+/* per-protocol mini bars in leaderboard */
+.proto-mini { display: flex; gap: 0.55rem; }
+.proto-mini > div { display: flex; flex-direction: column; gap: 0.22rem; min-width: 3.1rem; }
+.proto-mini .pl { font-family: var(--mono); font-size: 0.55rem; letter-spacing: 0.1em; text-transform: uppercase; color: var(--dim); }
+.proto-mini .pb { height: 3px; border-radius: 2px; background: #1b1f27; overflow: hidden; }
+.proto-mini .pb > i { display: block; height: 100%; width: 0; background: linear-gradient(90deg, #5b6ee0, #8f7bff); transition: width 1.1s var(--ease); }
+.proto-mini .pv { font-family: var(--mono); font-size: 0.64rem; color: var(--muted); }
+
+/* ---------------------------------------------------------- submit block */
+.submit {
+  border: 1px solid var(--line); border-radius: var(--radius); background: linear-gradient(180deg, var(--ink-2), #0b0d11);
+  padding: 1.7rem; position: relative; overflow: hidden;
+}
+.submit::before {
+  content: ""; position: absolute; inset: 0 auto 0 0; width: 3px;
+  background: linear-gradient(180deg, var(--accent), transparent);
+}
+.submit h2 { margin-top: 0; }
+.steps { margin: 1rem 0 0; padding-left: 1.3rem; color: var(--muted); display: grid; gap: 0.7rem; max-width: 62ch; }
+.steps b { color: var(--text); }
+code {
+  font-family: var(--mono); font-size: 0.84em; background: #0a0c10; border: 1px solid var(--line);
+  border-radius: 5px; padding: 0.06em 0.36em; color: #cdd3de;
+}
+pre {
+  margin: 1.1rem 0; padding: 1.1rem 1.25rem; border: 1px solid var(--line); border-radius: 10px;
+  background: #08090c; overflow-x: auto; font-size: 0.82rem; line-height: 1.62;
+}
+pre code { background: none; border: none; padding: 0; }
+
+/* -------------------------------------------------------------- metrics */
+.metrics { display: grid; grid-template-columns: repeat(auto-fit, minmax(220px, 1fr)); gap: 1rem; }
+.metric {
+  padding: 1.3rem; border: 1px solid var(--line); border-radius: var(--radius);
+  background: linear-gradient(180deg, var(--panel-2), var(--panel));
+  position: relative; overflow: hidden;
+}
+.metric::after {
+  content: ""; position: absolute; left: 1.3rem; right: 1.3rem; bottom: 0; height: 2px;
+  background: linear-gradient(90deg, var(--accent), transparent); opacity: 0.5;
+}
+.metric .label { font-family: var(--mono); font-size: 0.64rem; letter-spacing: 0.14em; text-transform: uppercase; color: var(--dim); margin-bottom: 0.4rem; }
+.metric .value { font-family: var(--mono); font-size: 1.7rem; font-variant-numeric: tabular-nums; font-weight: 600; }
+.metric .note { margin-top: 0.35rem; font-size: 0.78rem; color: var(--dim); }
+.metric .note code { font-size: 0.78em; }
+
+/* ------------------------------------------------------------------ footer */
+footer { border-top: 1px solid var(--line); padding: 2rem 0 3rem; color: var(--dim); font-size: 0.78rem; }
+footer .wrap { display: flex; flex-wrap: wrap; gap: 1.4rem; align-items: center; }
+footer .license { margin-left: auto; font-family: var(--mono); font-size: 0.7rem; }
+footer a { color: var(--muted); }
+
+/* ------------------------------------------------------------- reveal */
+.reveal { opacity: 1; transform: none; }
+.js .reveal { opacity: 0; transform: translateY(20px); transition: opacity 0.7s var(--ease), transform 0.7s var(--ease); will-change: opacity, transform; }
+.js .reveal.in { opacity: 1; transform: none; }
+.hero .anim { opacity: 0; transform: translateY(16px); animation: rise 0.8s var(--ease) forwards; }
+@keyframes rise { to { opacity: 1; transform: none; } }
+
+@media (max-width: 760px) {
+  .split { grid-template-columns: 1fr; }
+  .stats { gap: 0 1.3rem; }
+  .metrics { grid-template-columns: 1fr; }
+  .board th, .board td { padding: 0.6rem 0.55rem; font-size: 0.82rem; }
+  .bar { display: none; }
+  .nav-links a:not(.nav-cta):not([aria-current]) { display: none; }
+}
+
+@media (prefers-reduced-motion: reduce) {
+  * { animation: none !important; transition: none !important; }
+  .reveal { opacity: 1; transform: none; }
+  .hero .anim { opacity: 1; transform: none; }
+  tbody tr { opacity: 1; transform: none; }
+}