diff --git "a/code/benchmark_analysis.ipynb" "b/code/benchmark_analysis.ipynb"
--- "a/code/benchmark_analysis.ipynb"
+++ "b/code/benchmark_analysis.ipynb"
@@ -63,7 +63,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "id": "cell_001",
    "metadata": {
     "execution": {
@@ -182,7 +182,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": null,
    "id": "cell_003",
    "metadata": {
     "execution": {
@@ -192,18 +192,7 @@
      "shell.execute_reply": "2026-04-22T16:21:02.987571Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Total responses: 124,876\n",
-      "Unique participants: 1,290\n",
-      "Total stimuli (pairs): 9,800\n",
-      "Speakers: 100\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Load raw data\n",
     "responses = pd.read_csv(DATA_DIR / 'participant_responses.csv')\n",
@@ -218,7 +207,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": null,
    "id": "cell_004",
    "metadata": {
     "execution": {
@@ -228,16 +217,7 @@
      "shell.execute_reply": "2026-04-22T16:21:03.079195Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Full dataset:      124,876 responses from 1,290 participants\n",
-      "Qualified subset:  92,239 responses from 175 participants (reserved for individual-level baseline only)\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Participant metadata (computed but NOT used for primary benchmark target).\n",
     "# The preregistered participant filter (>=25 gold trials, >=60% accuracy) is used\n",
@@ -263,7 +243,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": null,
    "id": "cell_005",
    "metadata": {
     "execution": {
@@ -273,17 +253,7 @@
      "shell.execute_reply": "2026-04-22T16:21:03.128096Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Pairs with full data: 9800\n",
-      "Median responses per pair (full): 10\n",
-      "Min/max responses per pair: 8 / 92\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Compute P(same) from ALL 1,290 participants (primary benchmark target)\n",
     "def compute_psame(resp_df, stimuli_df):\n",
@@ -312,7 +282,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "id": "cell_006",
    "metadata": {
     "execution": {
@@ -322,121 +292,7 @@
      "shell.execute_reply": "2026-04-22T16:21:03.158257Z"
     }
    },
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "<div>\n",
-       "<style scoped>\n",
-       "    .dataframe tbody tr th:only-of-type {\n",
-       "        vertical-align: middle;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe tbody tr th {\n",
-       "        vertical-align: top;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe thead th {\n",
-       "        text-align: right;\n",
-       "    }\n",
-       "</style>\n",
-       "<table border=\"1\" class=\"dataframe\">\n",
-       "  <thead>\n",
-       "    <tr style=\"text-align: right;\">\n",
-       "      <th></th>\n",
-       "      <th>count</th>\n",
-       "      <th>mean_p_same</th>\n",
-       "      <th>std_p_same</th>\n",
-       "      <th>mean_n_resp</th>\n",
-       "      <th>description</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>stimuli_type</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>1</th>\n",
-       "      <td>100</td>\n",
-       "      <td>0.907</td>\n",
-       "      <td>0.072</td>\n",
-       "      <td>36.760</td>\n",
-       "      <td>Same audio (baseline)</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>2</th>\n",
-       "      <td>400</td>\n",
-       "      <td>0.689</td>\n",
-       "      <td>0.191</td>\n",
-       "      <td>39.970</td>\n",
-       "      <td>Same speaker, diff recording</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>3</th>\n",
-       "      <td>400</td>\n",
-       "      <td>0.593</td>\n",
-       "      <td>0.227</td>\n",
-       "      <td>10.060</td>\n",
-       "      <td>Voice clone (same speaker)</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>4</th>\n",
-       "      <td>400</td>\n",
-       "      <td>0.263</td>\n",
-       "      <td>0.169</td>\n",
-       "      <td>39.188</td>\n",
-       "      <td>Different speaker</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>5</th>\n",
-       "      <td>400</td>\n",
-       "      <td>0.274</td>\n",
-       "      <td>0.192</td>\n",
-       "      <td>10.068</td>\n",
-       "      <td>Different speaker clone</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>6</th>\n",
-       "      <td>8100</td>\n",
-       "      <td>0.381</td>\n",
-       "      <td>0.242</td>\n",
-       "      <td>10.060</td>\n",
-       "      <td>Interpolated/blended</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "              count  mean_p_same  std_p_same  mean_n_resp  \\\n",
-       "stimuli_type                                                \n",
-       "1               100        0.907       0.072       36.760   \n",
-       "2               400        0.689       0.191       39.970   \n",
-       "3               400        0.593       0.227       10.060   \n",
-       "4               400        0.263       0.169       39.188   \n",
-       "5               400        0.274       0.192       10.068   \n",
-       "6              8100        0.381       0.242       10.060   \n",
-       "\n",
-       "                               description  \n",
-       "stimuli_type                                \n",
-       "1                    Same audio (baseline)  \n",
-       "2             Same speaker, diff recording  \n",
-       "3               Voice clone (same speaker)  \n",
-       "4                        Different speaker  \n",
-       "5                  Different speaker clone  \n",
-       "6                     Interpolated/blended  "
-      ]
-     },
-     "execution_count": 5,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Stimulus type summary\n",
     "type_labels = {\n",
@@ -460,7 +316,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": null,
    "id": "cell_007",
    "metadata": {
     "execution": {
@@ -470,98 +326,7 @@
      "shell.execute_reply": "2026-04-22T16:21:03.183530Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Speaker demographics:\n",
-      "  Gender: {1: 50, 2: 50}\n",
-      "  Groups: 5 groups\n",
-      "  Age range: 1-2\n"
-     ]
-    },
-    {
-     "data": {
-      "text/html": [
-       "<div>\n",
-       "<style scoped>\n",
-       "    .dataframe tbody tr th:only-of-type {\n",
-       "        vertical-align: middle;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe tbody tr th {\n",
-       "        vertical-align: top;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe thead th {\n",
-       "        text-align: right;\n",
-       "    }\n",
-       "</style>\n",
-       "<table border=\"1\" class=\"dataframe\">\n",
-       "  <thead>\n",
-       "    <tr style=\"text-align: right;\">\n",
-       "      <th></th>\n",
-       "      <th>count</th>\n",
-       "      <th>mean</th>\n",
-       "      <th>std</th>\n",
-       "      <th>min</th>\n",
-       "      <th>25%</th>\n",
-       "      <th>50%</th>\n",
-       "      <th>75%</th>\n",
-       "      <th>max</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>gender</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>1</th>\n",
-       "      <td>50.0</td>\n",
-       "      <td>1.5</td>\n",
-       "      <td>0.5</td>\n",
-       "      <td>1.0</td>\n",
-       "      <td>1.0</td>\n",
-       "      <td>1.5</td>\n",
-       "      <td>2.0</td>\n",
-       "      <td>2.0</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>2</th>\n",
-       "      <td>50.0</td>\n",
-       "      <td>1.5</td>\n",
-       "      <td>0.5</td>\n",
-       "      <td>1.0</td>\n",
-       "      <td>1.0</td>\n",
-       "      <td>1.5</td>\n",
-       "      <td>2.0</td>\n",
-       "      <td>2.0</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "        count  mean  std  min  25%  50%  75%  max\n",
-       "gender                                           \n",
-       "1        50.0   1.5  0.5  1.0  1.0  1.5  2.0  2.0\n",
-       "2        50.0   1.5  0.5  1.0  1.0  1.5  2.0  2.0"
-      ]
-     },
-     "execution_count": 6,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Speaker demographics\n",
     "print('Speaker demographics:')\n",
@@ -596,7 +361,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": null,
    "id": "cell_008",
    "metadata": {
     "execution": {
@@ -606,16 +371,7 @@
      "shell.execute_reply": "2026-04-22T16:21:06.993670Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Split-half correlation (mean of 100 splits, all 1,290 participants): r = 0.5449\n",
-      "Spearman-Brown corrected reliability: 0.7054\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Inter-rater agreement via split-half reliability on ALL 1,290 participants\n",
     "N_SPLITS = 100\n",
@@ -641,7 +397,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": null,
    "id": "cell_009",
    "metadata": {
     "execution": {
@@ -651,18 +407,7 @@
      "shell.execute_reply": "2026-04-22T16:21:07.936924Z"
     }
    },
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 1000x500 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    }
-   ],
+   "outputs": [],
    "source": [
     "# P(same) distribution by stimulus type (full set, 1,290 participants)\n",
     "fig, ax = plt.subplots(figsize=(10, 5))\n",
@@ -688,20 +433,6 @@
     "plt.show()"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_009_interp",
-   "metadata": {},
-   "source": [
-    "### Section 1: Interpretation\n",
-    "\n",
-    "**Dataset scale.** The full dataset contains 124,876 judgments from 1,290 participants across 9,800 voice pairs spanning 100 speakers. Each pair has a median of 10 individual judgments (range 1-51).\n",
-    "\n",
-    "**P(same) by stimulus type.** The violin plot shows clear separation: Type 1 (identical audio) near ceiling (~0.95); Type 2 (same speaker, different recording) ~0.68; Type 3 (voice clone of same speaker) ~0.56; Types 4-5 (different speakers) ~0.24; Type 6 (blended voices) spans the full 0-1 range providing continuous identity-perception probes.\n",
-    "\n",
-    "**Inter-rater reliability.** Split-half Spearman-Brown reliability is $\\rho = 0.705$. This is the noise ceiling: a perfect predictor of the true (noise-free) P(same) would achieve Pearson $r \\approx \\sqrt{\\rho} = 0.84$ and $R^2 \\approx 0.705$ on this measurement. The reliability substantially below 1.0 reflects genuine population-level variability in how listeners perceive voice identity, not measurement error. Section 10's per-listener baseline makes this listener-variability quantitative."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "cell_010",
@@ -727,7 +458,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
+   "execution_count": null,
    "id": "cell_011",
    "metadata": {
     "execution": {
@@ -737,80 +468,7 @@
      "shell.execute_reply": "2026-04-22T16:21:42.947936Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded rawnet3: 9900 embeddings, dim=192\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded ecapa_tdnn: 9900 embeddings, dim=192\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded titanet: 9900 embeddings, dim=192\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded resemblyzer: 9900 embeddings, dim=256\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded xvector: 9900 embeddings, dim=512\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded wav2vec2: 9900 embeddings, dim=768\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded hubert: 9900 embeddings, dim=768\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded wavlm: 9900 embeddings, dim=768\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded whisper: 9900 embeddings, dim=512\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "  Loaded xlsr: 9900 embeddings, dim=1024\n",
-      "\n",
-      "Available models: ['rawnet3', 'ecapa_tdnn', 'titanet', 'resemblyzer', 'xvector', 'wav2vec2', 'hubert', 'wavlm', 'whisper', 'xlsr']\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Load all available embeddings\n",
     "MODEL_INFO = {\n",
@@ -844,7 +502,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": null,
    "id": "cell_012",
    "metadata": {
     "execution": {
@@ -854,72 +512,7 @@
      "shell.execute_reply": "2026-04-22T16:21:46.750883Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "rawnet3: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "ecapa_tdnn: 9800 / 9800 pairs with valid cosine similarity\n",
-      "titanet: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "resemblyzer: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "xvector: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "wav2vec2: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "hubert: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "wavlm: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "whisper: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "xlsr: 9800 / 9800 pairs with valid cosine similarity\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Compute cosine similarity for each model x each pair (vectorized)\n",
     "for model_name, emb_dict in embeddings.items():\n",
@@ -954,7 +547,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": null,
    "id": "cell_012b_bestlayer",
    "metadata": {
     "execution": {
@@ -964,24 +557,7 @@
      "shell.execute_reply": "2026-04-22T16:22:33.384012Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Nested-CV best-layer cosine similarity for SSL models...\n",
-      "[cache HIT] ssl_bestlayer_cosims: Per-fold layer selection for SSL models\n",
-      "  wav2vec2: best layer per fold = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0] (n_layers=13)\n",
-      "  hubert: best layer per fold = [1, 0, 1, 0, 0, 1, 0, 0, 1, 0] (n_layers=13)\n",
-      "  wavlm: best layer per fold = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0] (n_layers=13)\n",
-      "  xlsr: best layer per fold = [1, 1, 1, 1, 1, 1, 1, 1, 1, 1] (n_layers=25)\n",
-      "  whisper: best layer per fold = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3] (n_layers=7)\n",
-      "\n",
-      "SSL cosine similarity columns now use nested-CV best-layer selection.\n",
-      "Last-layer values preserved in `{model}_cosim_lastlayer` columns.\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# ============================================================\n",
     "# Nested speaker-level CV: best-layer cosine similarity for SSL models\n",
@@ -1083,7 +659,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
+   "execution_count": null,
    "id": "cell_013",
    "metadata": {
     "execution": {
@@ -1093,140 +669,7 @@
      "shell.execute_reply": "2026-04-22T16:22:33.398744Z"
     }
    },
-   "outputs": [
-    {
-     "data": {
-      "text/html": [
-       "<div>\n",
-       "<style scoped>\n",
-       "    .dataframe tbody tr th:only-of-type {\n",
-       "        vertical-align: middle;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe tbody tr th {\n",
-       "        vertical-align: top;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe thead th {\n",
-       "        text-align: right;\n",
-       "    }\n",
-       "</style>\n",
-       "<table border=\"1\" class=\"dataframe\">\n",
-       "  <thead>\n",
-       "    <tr style=\"text-align: right;\">\n",
-       "      <th></th>\n",
-       "      <th>Model</th>\n",
-       "      <th>Type</th>\n",
-       "      <th>Dim</th>\n",
-       "      <th>Training</th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>0</th>\n",
-       "      <td>rawnet3</td>\n",
-       "      <td>Supervised</td>\n",
-       "      <td>192</td>\n",
-       "      <td>AAM-Softmax on VoxCeleb1+2 (Jung et al., Inter...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>1</th>\n",
-       "      <td>ecapa_tdnn</td>\n",
-       "      <td>Supervised</td>\n",
-       "      <td>192</td>\n",
-       "      <td>AAM-Softmax on VoxCeleb1+2 (Desplanques et al....</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>2</th>\n",
-       "      <td>titanet</td>\n",
-       "      <td>Supervised</td>\n",
-       "      <td>192</td>\n",
-       "      <td>AAM-Softmax on VoxCeleb+Fisher+SWB (Koluguri e...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>3</th>\n",
-       "      <td>resemblyzer</td>\n",
-       "      <td>Supervised</td>\n",
-       "      <td>256</td>\n",
-       "      <td>GE2E loss, 3-layer LSTM (Wan et al., ICASSP 2018)</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>4</th>\n",
-       "      <td>xvector</td>\n",
-       "      <td>Supervised</td>\n",
-       "      <td>512</td>\n",
-       "      <td>Softmax on VoxCeleb1+2, TDNN (Snyder et al., I...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>5</th>\n",
-       "      <td>wav2vec2</td>\n",
-       "      <td>Self-supervised</td>\n",
-       "      <td>768</td>\n",
-       "      <td>Contrastive on LibriSpeech 960h (Baevski et al...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>6</th>\n",
-       "      <td>hubert</td>\n",
-       "      <td>Self-supervised</td>\n",
-       "      <td>768</td>\n",
-       "      <td>Masked prediction on LibriSpeech 960h (Hsu et ...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>7</th>\n",
-       "      <td>wavlm</td>\n",
-       "      <td>Self-supervised</td>\n",
-       "      <td>768</td>\n",
-       "      <td>Masked prediction + denoising on 94K hrs (Chen...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>8</th>\n",
-       "      <td>whisper</td>\n",
-       "      <td>Weakly supervised</td>\n",
-       "      <td>512</td>\n",
-       "      <td>Multitask ASR on 680K hrs web audio (Radford e...</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>9</th>\n",
-       "      <td>xlsr</td>\n",
-       "      <td>Self-supervised</td>\n",
-       "      <td>1024</td>\n",
-       "      <td>Contrastive on 436K hrs multilingual (Babu et ...</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "         Model               Type   Dim  \\\n",
-       "0      rawnet3         Supervised   192   \n",
-       "1   ecapa_tdnn         Supervised   192   \n",
-       "2      titanet         Supervised   192   \n",
-       "3  resemblyzer         Supervised   256   \n",
-       "4      xvector         Supervised   512   \n",
-       "5     wav2vec2    Self-supervised   768   \n",
-       "6       hubert    Self-supervised   768   \n",
-       "7        wavlm    Self-supervised   768   \n",
-       "8      whisper  Weakly supervised   512   \n",
-       "9         xlsr    Self-supervised  1024   \n",
-       "\n",
-       "                                            Training  \n",
-       "0  AAM-Softmax on VoxCeleb1+2 (Jung et al., Inter...  \n",
-       "1  AAM-Softmax on VoxCeleb1+2 (Desplanques et al....  \n",
-       "2  AAM-Softmax on VoxCeleb+Fisher+SWB (Koluguri e...  \n",
-       "3  GE2E loss, 3-layer LSTM (Wan et al., ICASSP 2018)  \n",
-       "4  Softmax on VoxCeleb1+2, TDNN (Snyder et al., I...  \n",
-       "5  Contrastive on LibriSpeech 960h (Baevski et al...  \n",
-       "6  Masked prediction on LibriSpeech 960h (Hsu et ...  \n",
-       "7  Masked prediction + denoising on 94K hrs (Chen...  \n",
-       "8  Multitask ASR on 680K hrs web audio (Radford e...  \n",
-       "9  Contrastive on 436K hrs multilingual (Babu et ...  "
-      ]
-     },
-     "execution_count": 12,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Model info table\n",
     "info_rows = []\n",
@@ -1257,6 +700,12 @@
    "source": [
     "# Inter-model correlation heatmap. Whisper is placed last (bottom-right) since\n",
     "# it is the weakly-supervised outlier.\n",
+    "DISPLAY_NAME = {'rawnet3': 'RawNet3', 'ecapa_tdnn': 'ECAPA-TDNN',\n",
+    "                'titanet': 'TitaNet', 'resemblyzer': 'resemblyzer',\n",
+    "                'xvector': 'x-vector', 'wav2vec2': 'wav2vec 2.0',\n",
+    "                'hubert': 'HuBERT', 'wavlm': 'WavLM',\n",
+    "                'whisper': 'Whisper', 'xlsr': 'XLS-R'}\n",
+    "\n",
     "if len(available_models) > 1:\n",
     "    plot_models = [m for m in available_models if m != 'whisper']\n",
     "    if 'whisper' in available_models:\n",
@@ -1266,12 +715,7 @@
     "    corr_matrix.columns = [DISPLAY_NAME[m] for m in plot_models]\n",
     "    corr_matrix.index = [DISPLAY_NAME[m] for m in plot_models]\n",
     "\n",
-    "    DISPLAY_NAME = {'rawnet3': 'RawNet3', 'ecapa_tdnn': 'ECAPA-TDNN',\n",
-    "                'titanet': 'TitaNet', 'resemblyzer': 'resemblyzer',\n",
-    "                'xvector': 'x-vector', 'wav2vec2': 'wav2vec 2.0',\n",
-    "                'hubert': 'HuBERT', 'wavlm': 'WavLM',\n",
-    "                'whisper': 'Whisper', 'xlsr': 'XLS-R'}\n",
-    "fig, ax = plt.subplots(figsize=(8, 6))\n",
+    "    fig, ax = plt.subplots(figsize=(8, 6))\n",
     "    sns.heatmap(corr_matrix, annot=True, fmt='.3f', cmap='RdYlBu_r',\n",
     "                vmin=0, vmax=1, ax=ax, square=True)\n",
     "    plt.tight_layout()\n",
@@ -1279,18 +723,6 @@
     "    plt.show()\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_014_interp",
-   "metadata": {},
-   "source": [
-    "### Section 2: Interpretation\n",
-    "\n",
-    "**Inter-model agreement.** With best-layer SSL cosine similarities, the inter-model correlation structure shifts: supervised and SSL models now correlate more strongly with each other than when SSL used last-layer. Supervised models still form the tightest cluster (pairwise r > 0.9 typical), but SSL models (wav2vec2, HuBERT, WavLM, XLS-R) now correlate with supervised models at r ~ 0.7-0.85 -- far above their last-layer correlations (~0.4-0.6). Whisper remains the outlier with weaker correlations to all other models. This pattern is consistent with SSL models encoding substantial speaker-relevant information in their early layers, which becomes visible under fair layer selection but is suppressed at the last layer.\n",
-    "\n",
-    "**Protocol note.** The cosine similarities computed here are the ones used throughout the rest of the benchmark. For supervised models they come from the model's native output; for SSL models they come from the nested-CV best-layer selection documented in Section 13."
-   ]
-  },
   {
    "cell_type": "markdown",
    "metadata": {},
@@ -1306,7 +738,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
+   "execution_count": null,
    "metadata": {
     "execution": {
      "iopub.execute_input": "2026-04-22T16:22:34.163892Z",
@@ -1315,18 +747,7 @@
      "shell.execute_reply": "2026-04-22T16:22:34.171417Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Pairs with metadata label: 1700 / 9800\n",
-      "  Metadata = same      (Types 1-3): 900\n",
-      "  Metadata = different (Types 4-5): 800\n",
-      "  Excluded             (Type 6):    8100\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# ============================================================\n",
     "# M1. Metadata labels (Type 6 excluded)\n",
@@ -1346,7 +767,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
+   "execution_count": null,
    "metadata": {
     "execution": {
      "iopub.execute_input": "2026-04-22T16:22:34.175974Z",
@@ -1355,30 +776,7 @@
      "shell.execute_reply": "2026-04-22T16:23:26.198051Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Nested-CV best-layer for SSL on the METADATA target (AUC-on-training)...\n",
-      "[cache MISS] ssl_bestlayer_cosims_meta: Per-fold layer selection on metadata target -- computing\n"
-     ]
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "[cache SAVED] ssl_bestlayer_cosims_meta -> ssl_bestlayer_cosims_meta.pkl\n",
-      "  wav2vec2: layers chosen = [0, 0, 0, 0, 0, 0, 0, 0, 3, 3] (of 13)\n",
-      "  hubert: layers chosen = [0, 0, 0, 0, 0, 0, 0, 0, 1, 1] (of 13)\n",
-      "  wavlm: layers chosen = [4, 4, 4, 4, 4, 4, 4, 4, 4, 4] (of 13)\n",
-      "  whisper: layers chosen = [3, 3, 3, 3, 3, 3, 3, 3, 3, 3] (of 7)\n",
-      "  xlsr: layers chosen = [5, 5, 5, 5, 5, 5, 5, 5, 5, 5] (of 25)\n",
-      "\n",
-      "Metadata cosines stored in df[f\"{m}_cosim_meta\"].\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# ============================================================\n",
     "# Per-fold best SSL layer selected on METADATA target (AUC on training speakers).\n",
@@ -1712,28 +1110,6 @@
     "print(f'\\nSpearman rank correlation (metadata AUC vs perception r): {rho:.4f}')\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### Part I.5: Interpretation\n",
-    "\n",
-    "**Metadata performance establishes baseline legitimacy.** On the VoxCeleb-comparable Set A (Type 2 same-speaker + Type 4 different-speaker), top supervised embeddings reach VoxCeleb-quality numbers: TitaNet EER = 1.6%, ECAPA-TDNN EER = 1.6%, RawNet3 EER = 1.9%. x-vector (11.1%) and resemblyzer (12.0%) are weaker but still strongly metadata-aligned (AUC \u2265 0.95). With per-task best-layer selection, SSL models reach EER = 26-32%; Whisper reaches 37.5%. This ordering matches speaker-verification literature: the models *do* work on the traditional task.\n",
-    "\n",
-    "**The dataset departs from metadata primarily through voice clones.** On Type 3 (same-speaker AI voice clones, labeled metadata=same), the crowd majority says **different** on 42% of pairs; median P(same) is 0.60. On other types, majority-vs-metadata disagreement is 0-19%. Overall, 27% of metadata=same pairs are judged as different by majority vote, and this divergence is concentrated in clones. This is the primary signal that the perception target is not a noisier version of metadata: it captures something different, specifically the cases where a cloned voice is acoustically the target speaker by ground truth yet is not heard that way.\n",
-    "\n",
-    "**Model rankings differ under metadata vs perception.** The Spearman correlation between AUC-on-metadata and Pearson-r-on-perception is 0.85 across the 10 models, so the rough ordering is preserved. But there are visible flips near the top: **resemblyzer is 5th on metadata AUC (0.949) yet 1st on perception r (0.804)**; TitaNet tops metadata but is 2nd on perception. The model that looks best on a VoxCeleb-style scoreboard is not the model best aligned with listeners.\n",
-    "\n",
-    "**For the paper.** These results support a three-beat narrative: (i) top supervised embeddings pass the traditional metadata test as expected; (ii) the benchmark's dataset departs from metadata primarily on voice clones, where 42% of metadata=same pairs are heard as different; (iii) model rankings under the perception target are not identical to metadata, so the two benchmarks are not interchangeable.\n"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "---"
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "part_II_divider",
@@ -1769,7 +1145,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 19,
+   "execution_count": null,
    "id": "cell_016",
    "metadata": {
     "execution": {
@@ -1779,26 +1155,7 @@
      "shell.execute_reply": "2026-04-22T16:23:27.709800Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Overall model-human alignment (full 1,290-participant dataset):\n",
-      "      model        model_type  pearson_r  spearman_rho  r_squared    n\n",
-      "resemblyzer        Supervised     0.6557        0.6682     0.4299 9800\n",
-      " ecapa_tdnn        Supervised     0.6456        0.6410     0.4168 9800\n",
-      "    titanet        Supervised     0.6381        0.6390     0.4072 9800\n",
-      "    rawnet3        Supervised     0.6171        0.6229     0.3809 9800\n",
-      "    xvector        Supervised     0.5973        0.6426     0.3568 9800\n",
-      "      wavlm   Self-supervised     0.5098        0.5121     0.2599 9800\n",
-      "   wav2vec2   Self-supervised     0.4714        0.4817     0.2222 9800\n",
-      "       xlsr   Self-supervised     0.4651        0.5080     0.2163 9800\n",
-      "     hubert   Self-supervised     0.4577        0.4773     0.2095 9800\n",
-      "    whisper Weakly supervised     0.2258        0.2698     0.0510 9800\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Overall correlation: each model's cosine similarity vs human P(same)\n",
     "def compute_correlations(df, model_name, target_col='p_same_full'):\n",
@@ -1893,7 +1250,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
+   "execution_count": null,
    "id": "cell_019",
    "metadata": {
     "execution": {
@@ -1903,167 +1260,7 @@
      "shell.execute_reply": "2026-04-22T16:23:32.988669Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Pearson r by stimulus type:\n"
-     ]
-    },
-    {
-     "data": {
-      "text/html": [
-       "<div>\n",
-       "<style scoped>\n",
-       "    .dataframe tbody tr th:only-of-type {\n",
-       "        vertical-align: middle;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe tbody tr th {\n",
-       "        vertical-align: top;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe thead th {\n",
-       "        text-align: right;\n",
-       "    }\n",
-       "</style>\n",
-       "<table border=\"1\" class=\"dataframe\">\n",
-       "  <thead>\n",
-       "    <tr style=\"text-align: right;\">\n",
-       "      <th></th>\n",
-       "      <th>Type 1</th>\n",
-       "      <th>Type 2</th>\n",
-       "      <th>Type 3</th>\n",
-       "      <th>Type 4</th>\n",
-       "      <th>Type 5</th>\n",
-       "      <th>Type 6</th>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>model</th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "      <th></th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>ecapa_tdnn</th>\n",
-       "      <td>0.348</td>\n",
-       "      <td>0.581</td>\n",
-       "      <td>0.453</td>\n",
-       "      <td>0.365</td>\n",
-       "      <td>0.306</td>\n",
-       "      <td>0.598</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>hubert</th>\n",
-       "      <td>0.173</td>\n",
-       "      <td>0.414</td>\n",
-       "      <td>0.366</td>\n",
-       "      <td>0.342</td>\n",
-       "      <td>0.320</td>\n",
-       "      <td>0.417</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>rawnet3</th>\n",
-       "      <td>0.180</td>\n",
-       "      <td>0.497</td>\n",
-       "      <td>0.336</td>\n",
-       "      <td>0.386</td>\n",
-       "      <td>0.314</td>\n",
-       "      <td>0.571</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>resemblyzer</th>\n",
-       "      <td>0.286</td>\n",
-       "      <td>0.650</td>\n",
-       "      <td>0.513</td>\n",
-       "      <td>0.520</td>\n",
-       "      <td>0.463</td>\n",
-       "      <td>0.607</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>titanet</th>\n",
-       "      <td>0.246</td>\n",
-       "      <td>0.612</td>\n",
-       "      <td>0.498</td>\n",
-       "      <td>0.409</td>\n",
-       "      <td>0.316</td>\n",
-       "      <td>0.585</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>wav2vec2</th>\n",
-       "      <td>0.247</td>\n",
-       "      <td>0.403</td>\n",
-       "      <td>0.360</td>\n",
-       "      <td>0.364</td>\n",
-       "      <td>0.305</td>\n",
-       "      <td>0.435</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>wavlm</th>\n",
-       "      <td>0.318</td>\n",
-       "      <td>0.453</td>\n",
-       "      <td>0.413</td>\n",
-       "      <td>0.392</td>\n",
-       "      <td>0.318</td>\n",
-       "      <td>0.467</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>whisper</th>\n",
-       "      <td>-0.084</td>\n",
-       "      <td>-0.007</td>\n",
-       "      <td>0.064</td>\n",
-       "      <td>0.063</td>\n",
-       "      <td>0.090</td>\n",
-       "      <td>0.216</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>xlsr</th>\n",
-       "      <td>0.304</td>\n",
-       "      <td>0.510</td>\n",
-       "      <td>0.414</td>\n",
-       "      <td>0.361</td>\n",
-       "      <td>0.325</td>\n",
-       "      <td>0.426</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>xvector</th>\n",
-       "      <td>0.232</td>\n",
-       "      <td>0.558</td>\n",
-       "      <td>0.451</td>\n",
-       "      <td>0.472</td>\n",
-       "      <td>0.370</td>\n",
-       "      <td>0.550</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "             Type 1  Type 2  Type 3  Type 4  Type 5  Type 6\n",
-       "model                                                      \n",
-       "ecapa_tdnn    0.348   0.581   0.453   0.365   0.306   0.598\n",
-       "hubert        0.173   0.414   0.366   0.342   0.320   0.417\n",
-       "rawnet3       0.180   0.497   0.336   0.386   0.314   0.571\n",
-       "resemblyzer   0.286   0.650   0.513   0.520   0.463   0.607\n",
-       "titanet       0.246   0.612   0.498   0.409   0.316   0.585\n",
-       "wav2vec2      0.247   0.403   0.360   0.364   0.305   0.435\n",
-       "wavlm         0.318   0.453   0.413   0.392   0.318   0.467\n",
-       "whisper      -0.084  -0.007   0.064   0.063   0.090   0.216\n",
-       "xlsr          0.304   0.510   0.414   0.361   0.325   0.426\n",
-       "xvector       0.232   0.558   0.451   0.472   0.370   0.550"
-      ]
-     },
-     "execution_count": 21,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Per-stimulus-type Pearson r\n",
     "type_corr_rows = []\n",
@@ -2085,27 +1282,6 @@
     "pivot"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_020_interp",
-   "metadata": {},
-   "source": [
-    "### Section 3: Interpretation\n",
-    "\n",
-    "**Supervised models hold a modest but robust lead over SSL models.** Full-dataset Pearson r values:\n",
-    "- Supervised: resemblyzer (r=0.656), ECAPA-TDNN (r=0.646), TitaNet (r=0.638), RawNet3 (r=0.617), x-vector (r=0.597).\n",
-    "- SSL (best-layer via nested CV): WavLM (r=0.510), wav2vec2 (r=0.471), XLS-R (r=0.465), HuBERT (r=0.458).\n",
-    "- Whisper (r=0.226) is far below the rest.\n",
-    "\n",
-    "The best supervised model exceeds the best SSL model by 0.15 in Pearson r. This is substantially smaller than the 0.38 gap that a naive last-layer-only SSL protocol would imply (Section 13).\n",
-    "\n",
-    "**Top supervised models cluster tightly.** Section 9's paired-bootstrap significance tests show resemblyzer vs ECAPA-TDNN is only marginally significant under Benjamini-Hochberg FDR control (diff=0.010, BH q=0.049). All other supervised-vs-supervised and supervised-vs-SSL differences are highly significant (BH q $\\ll$ 0.05).\n",
-    "\n",
-    "**Per-type breakdown.** All models perform worst on Type 1 (restricted P(same) variance near ceiling). Type 6 (blended voices) elicits the strongest correlations due to wide P(same) variance. Type 3 (clones) shows intermediate difficulty.\n",
-    "\n",
-    "**Noise ceiling.** The reliability-derived ceiling on R\u00b2 is 0.705. The best model (resemblyzer, R\u00b2 = 0.430) reaches **61% of the achievable ceiling**, leaving meaningful headroom that Section 10 attributes primarily to individual-listener variability rather than representational inadequacy."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "cell_021",
@@ -2311,7 +1487,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 24,
+   "execution_count": null,
    "id": "cell_024",
    "metadata": {
     "execution": {
@@ -2321,26 +1497,7 @@
      "shell.execute_reply": "2026-04-22T16:23:36.718999Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Human-Aligned Verification Results (Platt-scaled ECE):\n",
-      "      model              type    AUC  Acc@opt  Threshold  ECE_raw  ECE_calibrated\n",
-      "resemblyzer        Supervised 0.8722   0.7974     0.7627   0.3377          0.0266\n",
-      " ecapa_tdnn        Supervised 0.8621   0.7882     0.4385   0.0470          0.0180\n",
-      "    titanet        Supervised 0.8614   0.7824     0.4555   0.0575          0.0195\n",
-      "    xvector        Supervised 0.8573   0.7743     0.9533   0.5482          0.1357\n",
-      "    rawnet3        Supervised 0.8511   0.7824     0.5397   0.0862          0.0206\n",
-      "      wavlm   Self-supervised 0.7897   0.7041     0.6943   0.2716          0.0217\n",
-      "       xlsr   Self-supervised 0.7817   0.6962     0.9082   0.5017          0.0608\n",
-      "   wav2vec2   Self-supervised 0.7703   0.7107     0.7647   0.3357          0.0329\n",
-      "     hubert   Self-supervised 0.7653   0.7222     0.7635   0.3169          0.0346\n",
-      "    whisper Weakly supervised 0.6505   0.6430     0.9949   0.5944          0.0372\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Summary table: AUC, accuracy, raw ECE, calibrated ECE per model\n",
     "verification_rows = []\n",
@@ -2374,26 +1531,6 @@
     "print(verif_df.round(4).to_string(index=False))"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_024_interp",
-   "metadata": {},
-   "source": [
-    "### Section 4: Interpretation\n",
-    "\n",
-    "**AUC** (against full-dataset majority vote) places supervised models highest: resemblyzer 0.872, ECAPA-TDNN 0.862, TitaNet 0.861, x-vector 0.857, RawNet3 0.851. Best-layer SSL models cluster at AUC 0.77-0.79 (WavLM 0.790, XLS-R 0.782, wav2vec2 0.770, HuBERT 0.765). Whisper is at 0.651. The supervised lead in AUC is about 0.07-0.11 -- meaningful but not categorical.\n",
-    "\n",
-    "**Raw ECE is dominated by scale mismatch.** Models with similarity distributions concentrated in a narrow high-value range show high raw ECE regardless of representation quality, so raw ECE is not a meaningful comparison metric.\n",
-    "\n",
-    "**Platt-calibrated ECE reveals the real calibration picture:**\n",
-    "- Four supervised models achieve excellent calibration (ECE_cal 0.018-0.027): ECAPA-TDNN (0.018), TitaNet (0.020), RawNet3 (0.021), resemblyzer (0.027).\n",
-    "- Three SSL models match or beat most supervised models after calibration: WavLM (0.022), wav2vec2 (0.033), HuBERT (0.035), Whisper (0.037). The best-layer SSL protocol substantially improved SSL calibration relative to the last-layer default.\n",
-    "- XLS-R has mid-range calibration (0.061) despite strong AUC.\n",
-    "- **x-vector is the lone outlier** (ECE_cal = 0.136) -- an order of magnitude worse than the best-calibrated supervised models. Even Platt scaling cannot fully fix its confidence shape, suggesting a non-sigmoidal pathology in its cosine-similarity distribution.\n",
-    "\n",
-    "**Practical takeaway:** For deployment, always Platt-calibrate cosine similarities on held-out data. Most models then provide trustworthy probabilities; x-vector is the exception and may require isotonic regression or a more flexible calibrator."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "cell_025",
@@ -2417,7 +1554,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 25,
+   "execution_count": null,
    "id": "cell_026",
    "metadata": {
     "execution": {
@@ -2427,166 +1564,7 @@
      "shell.execute_reply": "2026-04-22T16:23:36.899533Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Unique reference speakers: 100\n",
-      "Unique comparison speakers: 100\n",
-      "\n",
-      "RDM entries: 400 (Types 4+5 off-diagonal, within-group)\n"
-     ]
-    },
-    {
-     "data": {
-      "text/html": [
-       "<div>\n",
-       "<style scoped>\n",
-       "    .dataframe tbody tr th:only-of-type {\n",
-       "        vertical-align: middle;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe tbody tr th {\n",
-       "        vertical-align: top;\n",
-       "    }\n",
-       "\n",
-       "    .dataframe thead th {\n",
-       "        text-align: right;\n",
-       "    }\n",
-       "</style>\n",
-       "<table border=\"1\" class=\"dataframe\">\n",
-       "  <thead>\n",
-       "    <tr style=\"text-align: right;\">\n",
-       "      <th></th>\n",
-       "      <th>reference</th>\n",
-       "      <th>comp_speaker</th>\n",
-       "      <th>p_same_full</th>\n",
-       "      <th>rawnet3_cosim</th>\n",
-       "      <th>ecapa_tdnn_cosim</th>\n",
-       "      <th>titanet_cosim</th>\n",
-       "      <th>resemblyzer_cosim</th>\n",
-       "      <th>xvector_cosim</th>\n",
-       "      <th>wav2vec2_cosim</th>\n",
-       "      <th>hubert_cosim</th>\n",
-       "      <th>wavlm_cosim</th>\n",
-       "      <th>whisper_cosim</th>\n",
-       "      <th>xlsr_cosim</th>\n",
-       "    </tr>\n",
-       "  </thead>\n",
-       "  <tbody>\n",
-       "    <tr>\n",
-       "      <th>0</th>\n",
-       "      <td>F01</td>\n",
-       "      <td>F02</td>\n",
-       "      <td>0.227778</td>\n",
-       "      <td>-0.096190</td>\n",
-       "      <td>-0.121267</td>\n",
-       "      <td>-0.246967</td>\n",
-       "      <td>0.515382</td>\n",
-       "      <td>0.903451</td>\n",
-       "      <td>0.649076</td>\n",
-       "      <td>0.630006</td>\n",
-       "      <td>0.627535</td>\n",
-       "      <td>0.992281</td>\n",
-       "      <td>0.876186</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>1</th>\n",
-       "      <td>F01</td>\n",
-       "      <td>F03</td>\n",
-       "      <td>0.092857</td>\n",
-       "      <td>0.047201</td>\n",
-       "      <td>-0.068396</td>\n",
-       "      <td>-0.175500</td>\n",
-       "      <td>0.642626</td>\n",
-       "      <td>0.915991</td>\n",
-       "      <td>0.772310</td>\n",
-       "      <td>0.754791</td>\n",
-       "      <td>0.704805</td>\n",
-       "      <td>0.992008</td>\n",
-       "      <td>0.897529</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>2</th>\n",
-       "      <td>F01</td>\n",
-       "      <td>F04</td>\n",
-       "      <td>0.101282</td>\n",
-       "      <td>0.146979</td>\n",
-       "      <td>0.213859</td>\n",
-       "      <td>0.094361</td>\n",
-       "      <td>0.662321</td>\n",
-       "      <td>0.911609</td>\n",
-       "      <td>0.781893</td>\n",
-       "      <td>0.759821</td>\n",
-       "      <td>0.722003</td>\n",
-       "      <td>0.994985</td>\n",
-       "      <td>0.914280</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>3</th>\n",
-       "      <td>F01</td>\n",
-       "      <td>F05</td>\n",
-       "      <td>0.217568</td>\n",
-       "      <td>0.269810</td>\n",
-       "      <td>0.160787</td>\n",
-       "      <td>0.132544</td>\n",
-       "      <td>0.661223</td>\n",
-       "      <td>0.912715</td>\n",
-       "      <td>0.814192</td>\n",
-       "      <td>0.777687</td>\n",
-       "      <td>0.812477</td>\n",
-       "      <td>0.994951</td>\n",
-       "      <td>0.932798</td>\n",
-       "    </tr>\n",
-       "    <tr>\n",
-       "      <th>4</th>\n",
-       "      <td>F02</td>\n",
-       "      <td>F01</td>\n",
-       "      <td>0.103191</td>\n",
-       "      <td>0.066491</td>\n",
-       "      <td>0.008276</td>\n",
-       "      <td>-0.045909</td>\n",
-       "      <td>0.522276</td>\n",
-       "      <td>0.918945</td>\n",
-       "      <td>0.592994</td>\n",
-       "      <td>0.681269</td>\n",
-       "      <td>0.548130</td>\n",
-       "      <td>0.990653</td>\n",
-       "      <td>0.830665</td>\n",
-       "    </tr>\n",
-       "  </tbody>\n",
-       "</table>\n",
-       "</div>"
-      ],
-      "text/plain": [
-       "  reference comp_speaker  p_same_full  rawnet3_cosim  ecapa_tdnn_cosim  \\\n",
-       "0       F01          F02     0.227778      -0.096190         -0.121267   \n",
-       "1       F01          F03     0.092857       0.047201         -0.068396   \n",
-       "2       F01          F04     0.101282       0.146979          0.213859   \n",
-       "3       F01          F05     0.217568       0.269810          0.160787   \n",
-       "4       F02          F01     0.103191       0.066491          0.008276   \n",
-       "\n",
-       "   titanet_cosim  resemblyzer_cosim  xvector_cosim  wav2vec2_cosim  \\\n",
-       "0      -0.246967           0.515382       0.903451        0.649076   \n",
-       "1      -0.175500           0.642626       0.915991        0.772310   \n",
-       "2       0.094361           0.662321       0.911609        0.781893   \n",
-       "3       0.132544           0.661223       0.912715        0.814192   \n",
-       "4      -0.045909           0.522276       0.918945        0.592994   \n",
-       "\n",
-       "   hubert_cosim  wavlm_cosim  whisper_cosim  xlsr_cosim  \n",
-       "0      0.630006     0.627535       0.992281    0.876186  \n",
-       "1      0.754791     0.704805       0.992008    0.897529  \n",
-       "2      0.759821     0.722003       0.994985    0.914280  \n",
-       "3      0.777687     0.812477       0.994951    0.932798  \n",
-       "4      0.681269     0.548130       0.990653    0.830665  "
-      ]
-     },
-     "execution_count": 25,
-     "metadata": {},
-     "output_type": "execute_result"
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Speaker-level RDM from Types 4+5 off-diagonal (different-speaker real + clone).\n",
     "# Type 6 (morphed) is excluded: at scale 0.5 a morph is neither A nor B, so the\n",
@@ -2627,7 +1605,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
+   "execution_count": null,
    "id": "cell_027",
    "metadata": {
     "execution": {
@@ -2637,26 +1615,7 @@
      "shell.execute_reply": "2026-04-22T16:24:02.346056Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Speaker-level RDM (Types 4+5, within-group different-speaker geometry):\n",
-      "      model              type  rsa_rho  mantel_p  n_pairs\n",
-      "resemblyzer        Supervised   0.5419    0.0000      400\n",
-      "    xvector        Supervised   0.4793    0.0000      400\n",
-      "      wavlm   Self-supervised   0.3923    0.0000      400\n",
-      "    rawnet3        Supervised   0.3856    0.0000      400\n",
-      "     hubert   Self-supervised   0.3829    0.0000      400\n",
-      "    titanet        Supervised   0.3792    0.0000      400\n",
-      "       xlsr   Self-supervised   0.3771    0.0000      400\n",
-      "   wav2vec2   Self-supervised   0.3569    0.0000      400\n",
-      " ecapa_tdnn        Supervised   0.3393    0.0000      400\n",
-      "    whisper Weakly supervised   0.1274    0.0092      400\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Mantel test: permute the model-side vector and compute empirical p-value.\n",
     "\n",
@@ -2694,28 +1653,6 @@
     "print(rsa_df.round(4).to_string(index=False))"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_027_interp",
-   "metadata": {},
-   "source": [
-    "### Section 5: Interpretation\n",
-    "\n",
-    "The speaker-level RDM (Types 4+5, 400 within-group off-diagonal entries) produces a ranking that **differs from pair-level Pearson r** in the middle tier:\n",
-    "\n",
-    "- Supervised: resemblyzer 0.535, x-vector 0.475, RawNet3 0.374, TitaNet 0.367, ECAPA-TDNN 0.329.\n",
-    "- SSL (best-layer): WavLM 0.389, HuBERT 0.382, XLS-R 0.375, wav2vec 2.0 0.355.\n",
-    "- Whisper 0.126 (Mantel p = 0.011; weakest model).\n",
-    "\n",
-    "**resemblyzer leads on both metrics**, but the rest of the ordering changes. **ECAPA-TDNN is the weakest supervised model on the speaker-level RDM** despite its second-highest Pearson r. Best-layer SSL models (0.36--0.39) sit between x-vector and the weaker supervised models, overlapping with them. This shows that models that rank individual voice pairs well are not necessarily models that preserve the speaker-to-speaker population map.\n",
-    "\n",
-    "**Whisper is a consistent outlier.** Its RDM correlation of 0.126 is significant (Mantel p = 0.011) but more than three times below every other model. This is consistent with its multitask ASR training discarding speaker-specific information at every layer.\n",
-    "\n",
-    "**Morph stimuli (Type 6) are not used in the RDM.** Aggregating morphs into a speaker-pair cell would conflate speaker-pair geometry with morph-trajectory geometry; a scale-0.5 morph is neither A nor B, so the \"comp speaker B\" label breaks down. Morph-trajectory behavior is examined separately via the Type 6 scale-by-scale curves (supplementary figure).\n",
-    "\n",
-    "**Scientific implication.** Pair-level and speaker-level metrics probe complementary properties. Strong pair-level alignment does not guarantee strong speaker-level geometry, and vice versa; benchmarks should report both."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "part_III_divider",
@@ -2749,7 +1686,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 27,
+   "execution_count": null,
    "id": "cell_030",
    "metadata": {
     "execution": {
@@ -2759,32 +1696,7 @@
      "shell.execute_reply": "2026-04-22T16:24:02.464293Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Cross-type transfer (REAL->SYNTH) with proper MSE/R^2 metrics:\n",
-      "Columns:\n",
-      "  R2_transfer:   R^2 on synthetic using linear map fit on REAL\n",
-      "  R2_oracle:     R^2 on synthetic using linear map fit on SYNTHETIC itself (oracle)\n",
-      "  R2_gap:        difference -- positive means real->synth transfer is LOSSY\n",
-      "  slope/intercept: reveal whether the real and synth mappings differ\n",
-      "\n",
-      "      model  R2_transfer  R2_oracle  R2_gap  slope_real  slope_synth  intercept_real  intercept_synth\n",
-      "    whisper      -0.1457     0.0455  0.1912     22.9004      10.8986        -22.2535         -10.4315\n",
-      "       xlsr       0.0033     0.1858  0.1825      4.9458       2.6242         -4.0101          -1.9717\n",
-      "    xvector       0.1408     0.3122  0.1714     10.5981       6.4585         -9.5866          -5.7224\n",
-      "   wav2vec2       0.0449     0.1921  0.1473      1.8367       1.1045         -0.8944          -0.4209\n",
-      "    rawnet3       0.1845     0.3301  0.1457      0.8226       0.6739          0.1046           0.0834\n",
-      "     hubert       0.0392     0.1790  0.1399      1.6520       0.9722         -0.7325          -0.3053\n",
-      "      wavlm       0.1314     0.2232  0.0918      1.6428       1.0092         -0.6885          -0.2836\n",
-      "    titanet       0.2658     0.3509  0.0851      0.8015       0.6903          0.1440           0.1196\n",
-      "resemblyzer       0.2974     0.3793  0.0819      2.4693       1.9108         -1.3698          -1.0148\n",
-      " ecapa_tdnn       0.2823     0.3618  0.0795      0.8154       0.7242          0.1548           0.1217\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Cross-stimulus-type transfer: train Platt-scaled predictor on real, test on synthetic\n",
     "# Use MSE and noise-ceiling-normalized R^2 on raw predicted values -- NOT Pearson r,\n",
@@ -2937,32 +1849,6 @@
     "    plt.show()\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_031_interp",
-   "metadata": {},
-   "source": [
-    "### Section 6: Interpretation\n",
-    "\n",
-    "**With best-layer SSL, real-to-synthetic transfer is no longer catastrophic.** The earlier finding of deeply negative R^2 for SSL models (when using last-layer) was substantially a layer-choice artifact. With nested-CV best-layer SSL, R^2_transfer is near-zero or weakly positive for most models:\n",
-    "- Supervised: positive R^2_transfer (0.14-0.30); oracle R^2 is 0.31-0.38; R^2 gap ~0.08-0.17.\n",
-    "- SSL (best-layer): near-zero R^2_transfer (0.003-0.13); oracle R^2 is 0.18-0.22; R^2 gap ~0.09-0.18.\n",
-    "- Whisper is the clear exception: R^2_transfer = -0.15, R^2_oracle = 0.05, gap = 0.19.\n",
-    "\n",
-    "**The slope differences persist but vary in magnitude.** For most models, the real-speech calibration slope is 1.1-2.1x steeper than the synthetic-speech slope. Representative values:\n",
-    "- wav2vec2: slope_real=1.84, slope_synth=1.10 (ratio 1.66)\n",
-    "- HuBERT: slope_real=1.65, slope_synth=0.97 (ratio 1.70)\n",
-    "- Whisper: slope_real=22.9, slope_synth=10.9 (ratio 2.10; large magnitudes reflect Whisper's narrow cosim range)\n",
-    "- resemblyzer: slope_real=2.47, slope_synth=1.91 (ratio 1.29)\n",
-    "- ECAPA-TDNN: slope_real=0.82, slope_synth=0.72 (ratio 1.13)\n",
-    "\n",
-    "Supervised models (ECAPA-TDNN, TitaNet, RawNet3) have slope ratios closest to 1.0, meaning their real-vs-synthetic calibration curves are most similar.\n",
-    "\n",
-    "**Scientific implication.** Voice cloning and interpolation produce systematic distribution shifts in cosine similarity even when the representation captures identity well. A benchmark that evaluates only on real speech mis-estimates synthetic-speech calibration. This is now a calibration issue rather than a representation failure.\n",
-    "\n",
-    "**Paper framing.** The real\u2192synth transfer finding is a **methodological warning**: regardless of architecture, a model deployed for voice-clone evaluation needs its similarity-to-probability mapping calibrated on synthetic data, not just real data. The effect is largest for weaker models (Whisper) but present for all."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "cell_032",
@@ -2986,7 +1872,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 29,
+   "execution_count": null,
    "id": "cell_033",
    "metadata": {
     "execution": {
@@ -2996,17 +1882,7 @@
      "shell.execute_reply": "2026-04-22T16:24:03.134575Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Mean human entropy: 0.7537\n",
-      "Median human entropy: 0.8813\n",
-      "Pairs with high disagreement (entropy > 0.9): 3167\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Human disagreement: entropy of vote distribution per pair\n",
     "def binary_entropy(p):\n",
@@ -3025,7 +1901,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 30,
+   "execution_count": null,
    "id": "cell_034",
    "metadata": {
     "execution": {
@@ -3035,26 +1911,7 @@
      "shell.execute_reply": "2026-04-22T16:24:03.235070Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Model confidence vs human agreement:\n",
-      "      model  r_confidence_agreement   p\n",
-      "resemblyzer                  0.3440 0.0\n",
-      "    xvector                  0.3033 0.0\n",
-      " ecapa_tdnn                  0.2853 0.0\n",
-      "    titanet                  0.2764 0.0\n",
-      "    rawnet3                  0.2755 0.0\n",
-      "      wavlm                  0.2415 0.0\n",
-      "     hubert                  0.2185 0.0\n",
-      "   wav2vec2                  0.2167 0.0\n",
-      "       xlsr                  0.2094 0.0\n",
-      "    whisper                  0.1082 0.0\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Can models predict human disagreement?\n",
     "disagree_results = []\n",
@@ -3087,7 +1944,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 31,
+   "execution_count": null,
    "id": "cell_035",
    "metadata": {
     "execution": {
@@ -3097,33 +1954,7 @@
      "shell.execute_reply": "2026-04-22T16:24:03.252797Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "High-disagreement pairs: 3167\n",
-      "\n",
-      "By stimulus type:\n",
-      "stimuli_type\n",
-      "1       2\n",
-      "2     149\n",
-      "3     177\n",
-      "4     110\n",
-      "5      99\n",
-      "6    2630\n",
-      "Name: count, dtype: int64\n",
-      "\n",
-      "Proportion of each type that is high-disagreement:\n",
-      "  Type 1: 2/100 = 2.0%\n",
-      "  Type 2: 149/400 = 37.2%\n",
-      "  Type 3: 177/400 = 44.2%\n",
-      "  Type 4: 110/400 = 27.5%\n",
-      "  Type 5: 99/400 = 24.8%\n",
-      "  Type 6: 2630/8100 = 32.5%\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Characterize high-disagreement pairs\n",
     "high_disagree = df[df['human_entropy'] > 0.9].copy()\n",
@@ -3137,25 +1968,6 @@
     "    print(f'  Type {stype}: {high}/{total} = {high/total:.1%}')"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_035_interp",
-   "metadata": {},
-   "source": [
-    "### Section 7: Interpretation\n",
-    "\n",
-    "**Substantial human disagreement exists.** Mean pair-level entropy is 0.75 and 32% of pairs (3,167 of 9,800) have entropy > 0.9 (near-maximal disagreement).\n",
-    "\n",
-    "**High-disagreement pairs are concentrated in Type 3 (clones, 44%) and Type 6 (blends, 33%)** -- exactly the stimulus types designed to probe identity boundaries. Type 1 (same audio) has only 2% high-disagreement pairs, confirming it is a reliable baseline.\n",
-    "\n",
-    "**Model confidence modestly predicts human agreement.** Correlations between model confidence (distance from optimal threshold) and human agreement (1 - entropy):\n",
-    "- Supervised: 0.28-0.34 (resemblyzer 0.34, x-vector 0.30, ECAPA-TDNN 0.29, TitaNet 0.28, RawNet3 0.28).\n",
-    "- SSL (best-layer): 0.21-0.24 (WavLM 0.24, HuBERT 0.22, wav2vec2 0.22, XLS-R 0.21).\n",
-    "- Whisper: 0.11.\n",
-    "\n",
-    "**Practical implication.** No current model reliably flags pairs humans will find ambiguous -- even the best model achieves only r=0.34. For confidence-aware verification systems, additional signals (e.g., stimulus-level features, listener priors, context) would be needed to anticipate disagreement zones."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "part_IV_divider",
@@ -3196,7 +2008,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 32,
+   "execution_count": null,
    "id": "cell_037",
    "metadata": {
     "execution": {
@@ -3206,17 +2018,7 @@
      "shell.execute_reply": "2026-04-22T16:24:03.267179Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Mahalanobis metric learning with 10-fold speaker-level CV...\n",
-      "(This may take a few minutes for high-dimensional models)\n",
-      "\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Learn per-dimension weights (diagonal Mahalanobis) via cross-validation\n",
     "from scipy.special import expit as sigmoid\n",
@@ -3439,26 +2241,6 @@
     "    plt.show()\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "cell_039_interp",
-   "metadata": {},
-   "source": [
-    "### Section 8: Interpretation\n",
-    "\n",
-    "**Protocol.** Mahalanobis learning uses each model's **main-benchmark embedding**: supervised models use their native output (192-512 dim); SSL models use the **per-fold best-layer embedding** (consistent with Section 2 and all downstream analyses). High-dimensional embeddings (> 256 dim) are reduced to 50 dim via PCA to keep parameter count tractable.\n",
-    "\n",
-    "**For supervised models, Mahalanobis does NOT improve over isotropic distance** (deltas near zero or negative: resemblyzer -0.032, ECAPA-TDNN -0.014, RawNet3 -0.005, TitaNet -0.002, x-vector +0.014). Supervised speaker embeddings are already approximately isotropically aligned with human identity perception.\n",
-    "\n",
-    "**For best-layer SSL embeddings, Mahalanobis provides modest but consistent improvement** (XLS-R +0.035, HuBERT +0.032, wav2vec2 +0.028, WavLM +0.019, Whisper +0.104). SSL representations, even at their best-aligned layer, contain identity-relevant information distributed non-uniformly across dimensions. Learned reweighting helps, but not enough to close the supervised-vs-SSL gap (SSL Mahalanobis R^2 remains 0.18-0.20 vs supervised 0.32-0.40).\n",
-    "\n",
-    "**Whisper's anomalously large Mahalanobis gain** (+0.104) is notable: its representations are dominated by ASR-relevant features that are largely orthogonal to speaker identity, so a learned metric can preferentially weight the small subset of dimensions carrying speaker information. Its absolute ceiling remains low (R^2=0.13).\n",
-    "\n",
-    "**Practical recommendation.** For supervised speaker embeddings, simple cosine similarity is optimal. For best-layer SSL representations, a learned diagonal metric yields a small non-trivial improvement but does not close the supervised gap.\n",
-    "\n",
-    "**Limitation.** We use only a diagonal (per-dimension) Mahalanobis. A full-matrix metric could capture cross-dimension interactions but would need more data or stronger regularization."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "part_V_divider",
@@ -3602,22 +2384,6 @@
     "          f'[{ci_low[top_idx, j]:+.4f}, {ci_high[top_idx, j]:+.4f}]  q_BH = {q_matrix[top_idx, j]:.4f}')\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "sec09_interp",
-   "metadata": {},
-   "source": [
-    "### Section 9: Interpretation\n",
-    "\n",
-    "**Of 45 unique pairwise comparisons between the 10 models, 43 are statistically significant under Benjamini-Hochberg false-discovery-rate control at $\\alpha = 0.05$.** The two non-significant comparisons sit within the SSL middle tier (wav2vec 2.0 vs XLS-R, BH q=0.367; HuBERT vs XLS-R, BH q=0.288). The resemblyzer vs ECAPA-TDNN comparison is significant but only marginally so (BH q=0.049), placing those two on adjacent boundaries of the supervised top tier rather than as a clean joint tier.\n",
-    "\n",
-    "**Supervised-vs-SSL separations are all highly significant** (diff > 0.08, BH q $\\ll$ 0.05), substantially smaller in absolute magnitude than the 0.38 gap a last-layer-only SSL protocol would produce.\n",
-    "\n",
-    "**Within the SSL cluster**, WavLM (r=0.510) is significantly ahead of every other SSL model (BH q < 0.05). The remaining three SSL models (wav2vec 2.0, XLS-R, HuBERT) overlap (XLS-R is not significantly different from wav2vec 2.0 or HuBERT under BH).\n",
-    "\n",
-    "**For the paper:** the three-tier structure (supervised > SSL > Whisper) is supported, with within-tier ties documented in the BH q matrix. Include the pairwise significance heatmap as a supplementary figure."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "sec10_header",
@@ -3784,34 +2550,6 @@
     "        print(f'  {model_name:15s}  acc={acc:.4f}  beats {pct:.1f}% of individual humans')\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "sec10_interp",
-   "metadata": {},
-   "source": [
-    "### Section 10: Interpretation\n",
-    "\n",
-    "**What the numbers say.** Of 1,290 total participants, 457 completed at least 25 trials (the minimum needed for a reliable per-participant accuracy estimate). Their mean agreement with the leave-one-out population consensus is 0.73 (std 0.09, range 0.34-0.94). The wide range reflects that this set includes listeners who completed enough trials for a reliable estimate but did not pass the PNAS-preregistered quality filter, including some low-agreement individuals.\n",
-    "\n",
-    "Model agreement (at each model's optimal threshold):\n",
-    "- **resemblyzer (Acc=0.80) beats 82% of individual humans.**\n",
-    "- **ECAPA-TDNN (0.79) beats 78%** of individual humans.\n",
-    "- TitaNet (0.78) beats 76%; RawNet3 (0.78) beats 76%; x-vector (0.77) beats 72%.\n",
-    "- Best-layer SSL: HuBERT (0.72) beats 37%, wav2vec2 (0.71) beats 31%, WavLM (0.70) beats 28%, XLS-R (0.70) beats 25%.\n",
-    "- Whisper (0.64) beats only 14% of individual humans.\n",
-    "\n",
-    "**Headline statement for the paper:** State-of-the-art supervised speaker embeddings agree with the population-consensus voice-identity label more often than **76-82% of individual listeners** who completed enough trials. Best-layer SSL models reach **25-37%** of the human distribution. Even Whisper, the weakest benchmarked model, matches some 14% of listeners.\n",
-    "\n",
-    "**What we do NOT claim.** Models are NOT \"more accurate than humans\" in an absolute sense; the population consensus itself is probabilistic for ambiguous pairs (voice clones, mid-blends). The claim is about consensus alignment: models rank and decide in closer agreement with the population than the majority of individuals do.\n",
-    "\n",
-    "**Connection to the noise ceiling.** The 61%-of-ceiling gap in Section 3 and the individual-human baseline describe the same listener-variability phenomenon from two angles: individual humans also fall short of the population consensus (mean individual agreement = 0.73 vs. model-to-consensus agreement = 0.80 for the best model). Closing the remaining gap requires modeling listener heterogeneity (e.g., listener-conditioned embeddings, personalized identity thresholds), not better speaker representations per se.\n",
-    "\n",
-    "**Limitations.**\n",
-    "1. The consensus is drawn from a specific US-English crowdworker pool; a different population would produce a different consensus.\n",
-    "2. The >=25-trials filter is a statistical requirement for reliable per-participant accuracy estimation, not a quality filter; it leaves a listener distribution that includes both high-engagement and low-engagement responders.\n",
-    "3. Model thresholds were tuned on the full majority-vote labels (Section 4), giving models a mild tuning advantage over individual humans. A fully nested protocol would slightly reduce the model percentiles but not change the ordering."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "sec11_header",
@@ -3827,7 +2565,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 37,
+   "execution_count": null,
    "id": "sec11_code",
    "metadata": {
     "execution": {
@@ -3837,31 +2575,7 @@
      "shell.execute_reply": "2026-04-22T16:24:20.160647Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Pairs excluding Type 6: 1700\n",
-      "Binary-classifiable (exclude ties): 1594\n",
-      "\n",
-      "Type 6 ablation: metrics with and without blended voices:\n",
-      "      model  r (Types 1-6)  r (Types 1-5 only)  r delta  AUC (Types 1-6)  AUC (Types 1-5 only)  AUC delta\n",
-      "resemblyzer         0.6557              0.8036   0.1479           0.8722                0.9293     0.0571\n",
-      "    titanet         0.6381              0.7944   0.1563           0.8614                0.9213     0.0599\n",
-      " ecapa_tdnn         0.6456              0.7910   0.1455           0.8621                0.9180     0.0558\n",
-      "    rawnet3         0.6171              0.7736   0.1565           0.8511                0.9128     0.0616\n",
-      "    xvector         0.5973              0.7618   0.1645           0.8573                0.9181     0.0607\n",
-      "      wavlm         0.5098              0.6151   0.1053           0.7897                0.8220     0.0323\n",
-      "       xlsr         0.4651              0.5911   0.1260           0.7817                0.8110     0.0293\n",
-      "   wav2vec2         0.4714              0.5731   0.1016           0.7703                0.7996     0.0293\n",
-      "     hubert         0.4577              0.5685   0.1108           0.7653                0.7989     0.0335\n",
-      "    whisper         0.2258              0.2030  -0.0228           0.6505                0.6485    -0.0021\n",
-      "\n",
-      "Spearman rank correlation between full-set and Type-1-5-only rankings: 0.9758\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Recompute Pearson r and AUC on Types 1-5 only (exclude Type 6 blends)\n",
     "\n",
@@ -3913,20 +2627,6 @@
     "print(f'\\nSpearman rank correlation between full-set and Type-1-5-only rankings: {rank_corr:.4f}')"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "sec11_interp",
-   "metadata": {},
-   "source": [
-    "### Section 11: Interpretation\n",
-    "\n",
-    "**Rankings are essentially preserved** when Type 6 is excluded: Spearman rank correlation between full-set and Types-1-5-only rankings is 0.976. Absolute Pearson r *increases* on the Types-1-5 subset for every model except Whisper (deltas typically +0.10 to +0.17). Type 6 has high within-pair noise (many blend stimuli at each ratio, judged by few participants each), which drags down pair-level correlations. The blended stimuli make the benchmark harder, not easier.\n",
-    "\n",
-    "**Whisper is the one exception**: its Type-6-excluded r (0.203) is *below* its full-set r (0.226) -- the only model whose performance worsens without blends. This reflects that Whisper's small residual identity signal is most visible over Type 6's wide P(same) range; on the narrower range of Types 1-5, what little signal it has is buried in noise. This is another indicator of Whisper's weak speaker representation rather than a property of the benchmark.\n",
-    "\n",
-    "**Implication for the paper.** The benchmark's conclusions hold without Type 6. Model rankings reflect genuine voice-identity perception alignment rather than an artifact of the blended-voice stimulus type. The Type-6-excluded table belongs in the supplementary material as a robustness check."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "sec12_header",
@@ -4017,27 +2717,6 @@
     "plt.show()"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "sec12_interp",
-   "metadata": {},
-   "source": [
-    "### Section 12: Interpretation\n",
-    "\n",
-    "**Gender disparities are small for all benchmark-competitive models.** Max gender gap |r_female - r_male|:\n",
-    "- Supervised: 0.001-0.014 (ECAPA-TDNN 0.001, resemblyzer 0.002, TitaNet 0.011, RawNet3 0.013, x-vector 0.014). Effectively gender-neutral.\n",
-    "- SSL (best-layer): 0.003-0.014 (WavLM 0.003, HuBERT 0.003, wav2vec2 0.014, XLS-R 0.014). Also effectively neutral.\n",
-    "- **Whisper** shows a substantially larger disparity (0.122) -- male speakers are far better predicted than female. This is consistent with Whisper's training corpus having known gender imbalances.\n",
-    "\n",
-    "**Sociophonetic group disparities (max-min across 5 groups) cluster around 0.11-0.17 for all models.** Groups 1-3 tend to be best-predicted, Groups 4-5 worst. Because the pattern holds across supervised and SSL families, it likely reflects stimulus-level variability between groups rather than model-specific bias.\n",
-    "\n",
-    "**Best-layer SSL substantially improved gender fairness** compared to last-layer SSL. Earlier last-layer numbers showed SSL gender gaps of 0.05-0.08; with best-layer selection these drop to 0.003-0.014. The suppression of speaker-relevant features in deeper SSL layers was partially gender-correlated; early layers encode identity more uniformly across gender.\n",
-    "\n",
-    "**Caveats.** With 20 speakers per group and 50 per gender, statistical power for small subgroup-difference tests is modest. We report descriptive disparities, not inferential tests. Fairness certification would require a larger and more demographically stratified speaker pool.\n",
-    "\n",
-    "**For the paper.** The gender-equitable result holds for all models except Whisper. Sociophonetic disparities are a limitation worth flagging but not model-specific."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "sec13_header",
@@ -4058,7 +2737,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 39,
+   "execution_count": null,
    "id": "sec13_code",
    "metadata": {
     "execution": {
@@ -4068,50 +2747,7 @@
      "shell.execute_reply": "2026-04-22T16:24:21.499697Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "Reused wav2vec2_layers from Section 2: 9900 clips, shape (13, 768)\n",
-      "Reused hubert_layers from Section 2: 9900 clips, shape (13, 768)\n",
-      "Reused wavlm_layers from Section 2: 9900 clips, shape (13, 768)\n",
-      "Reused xlsr_layers from Section 2: 9900 clips, shape (25, 1024)\n",
-      "Reused whisper_layers from Section 2: 9900 clips, shape (7, 512)\n",
-      "[cache HIT] layer_alignment: Per-layer Pearson r for SSL models\n",
-      "\n",
-      "wav2vec2 (13 layers):\n",
-      "  per-layer r: [np.float64(0.471), np.float64(0.457), np.float64(0.459), np.float64(0.436), np.float64(0.413), np.float64(0.396), np.float64(0.382), np.float64(0.363), np.float64(0.369), np.float64(0.391), np.float64(0.387), np.float64(0.21), np.float64(0.223)]\n",
-      "  best layer: 0 (r=0.4714)\n",
-      "  last layer (L=12, naive out-of-the-box baseline; NOT used in main benchmark): r=0.2232\n",
-      "  improvement from layer selection: +0.2482\n",
-      "\n",
-      "hubert (13 layers):\n",
-      "  per-layer r: [np.float64(0.48), np.float64(0.478), np.float64(0.407), np.float64(0.384), np.float64(0.359), np.float64(0.351), np.float64(0.344), np.float64(0.323), np.float64(0.307), np.float64(0.301), np.float64(0.308), np.float64(0.333), np.float64(0.269)]\n",
-      "  best layer: 0 (r=0.4795)\n",
-      "  last layer (L=12, naive out-of-the-box baseline; NOT used in main benchmark): r=0.2692\n",
-      "  improvement from layer selection: +0.2103\n",
-      "\n",
-      "wavlm (13 layers):\n",
-      "  per-layer r: [np.float64(0.51), np.float64(0.491), np.float64(0.456), np.float64(0.433), np.float64(0.432), np.float64(0.412), np.float64(0.384), np.float64(0.351), np.float64(0.3), np.float64(0.269), np.float64(0.27), np.float64(0.148), np.float64(0.227)]\n",
-      "  best layer: 0 (r=0.5098)\n",
-      "  last layer (L=12, naive out-of-the-box baseline; NOT used in main benchmark): r=0.2271\n",
-      "  improvement from layer selection: +0.2827\n",
-      "\n",
-      "xlsr (25 layers):\n",
-      "  per-layer r: [np.float64(0.459), np.float64(0.465), np.float64(0.445), np.float64(0.438), np.float64(0.436), np.float64(0.455), np.float64(0.41), np.float64(0.379), np.float64(0.376), np.float64(0.366), np.float64(0.347), np.float64(0.328), np.float64(0.324), np.float64(0.32), np.float64(0.325), np.float64(0.328), np.float64(0.335), np.float64(0.361), np.float64(0.4), np.float64(0.408), np.float64(0.253), np.float64(0.293), np.float64(0.031), np.float64(0.03), np.float64(0.027)]\n",
-      "  best layer: 1 (r=0.4651)\n",
-      "  last layer (L=24, naive out-of-the-box baseline; NOT used in main benchmark): r=0.0270\n",
-      "  improvement from layer selection: +0.4381\n",
-      "\n",
-      "whisper (7 layers):\n",
-      "  per-layer r: [np.float64(0.094), np.float64(0.104), np.float64(0.148), np.float64(0.226), np.float64(0.113), np.float64(0.069), np.float64(0.06)]\n",
-      "  best layer: 3 (r=0.2258)\n",
-      "  last layer (L=6, naive out-of-the-box baseline; NOT used in main benchmark): r=0.0604\n",
-      "  improvement from layer selection: +0.1655\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Layer-wise alignment for SSL models (reuses layer_embs from Section 2; cached)\n",
     "\n",
@@ -4184,7 +2820,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 40,
+   "execution_count": null,
    "id": "sec13_plot",
    "metadata": {
     "execution": {
@@ -4194,36 +2830,7 @@
      "shell.execute_reply": "2026-04-22T16:24:21.982008Z"
     }
    },
-   "outputs": [
-    {
-     "data": {
-      "image/png": "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",
-      "text/plain": [
-       "<Figure size 900x500 with 1 Axes>"
-      ]
-     },
-     "metadata": {},
-     "output_type": "display_data"
-    },
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "\n",
-      "SSL last-layer vs best-layer summary:\n",
-      "   model  n_layers  r_last_layer (used in main)  r_best_layer  best_layer_index  best_layer_norm  improvement\n",
-      "wav2vec2        13                       0.2232        0.4714                 0            0.000       0.2482\n",
-      "  hubert        13                       0.2692        0.4795                 0            0.000       0.2103\n",
-      "   wavlm        13                       0.2271        0.5098                 0            0.000       0.2827\n",
-      "    xlsr        25                       0.0270        0.4651                 1            0.042       0.4381\n",
-      " whisper         7                       0.0604        0.2258                 3            0.500       0.1655\n",
-      "\n",
-      "Best supervised model: r = 0.6557\n",
-      "Best SSL (last layer):      r = 0.2692 (gap to supervised: 0.3864)\n",
-      "Best SSL (best layer):      r = 0.5098 (gap to supervised: 0.1458)\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# Plot: Pearson r vs normalized layer index for the four SSL models.\n",
     "# Whisper is excluded -- it is weakly supervised, not SSL.\n",
@@ -4293,28 +2900,6 @@
     "print(f'Best SSL (best layer):      r = {best_ssl_best_r:.4f} (gap to supervised: {best_sup_r - best_ssl_best_r:.4f})')\n"
    ]
   },
-  {
-   "cell_type": "markdown",
-   "id": "sec13_interp",
-   "metadata": {},
-   "source": [
-    "### Section 13: Interpretation\n",
-    "\n",
-    "**Per-layer alignment (Pearson r with P(same)) reveals consistent patterns:**\n",
-    "- **wav2vec2, HuBERT, WavLM, XLS-R**: speaker information peaks at or near layer 0 (CNN feature-projection output) or layer 1 (first transformer block), and declines monotonically in deeper layers. By the final layer, alignment drops substantially (e.g., WavLM: 0.50 at layer 0 -> 0.22 at layer 12; XLS-R: 0.46 at layer 1 -> 0.03 at layer 24).\n",
-    "- **Whisper is different.** Its alignment peaks at layer 3 (middle of the 7-layer encoder) at only r=0.22, and both earlier and later layers are weaker. This reflects Whisper's multitask ASR training: speaker identity is not preserved at any depth.\n",
-    "- **XLS-R's last 3 layers (22-24) are near-chance for speaker identity.** These layers specialize toward multilingual phonetic discrimination, not speaker features. Using XLS-R's last layer (as the out-of-the-box default would) gives r = 0.03 -- effectively useless for speaker verification. The nested-CV best-layer protocol (Section 2) correctly identifies layer 1 instead.\n",
-    "\n",
-    "**Diagnostic for the main benchmark:** The nested-CV layer selection consistently picks layers 0 or 1 for wav2vec2/HuBERT/WavLM/XLS-R and layer 3 for Whisper. These selections are stable across the 10 CV folds (verified by the `best_layer_per_fold` dictionary computed in Section 2), meaning the layer choice is not fragile to which speakers are held out.\n",
-    "\n",
-    "**Why last-layer would be misleading.** If the benchmark reported only last-layer SSL numbers (the naive default), it would:\n",
-    "- Undersell wav2vec2 (0.22 instead of 0.46), HuBERT (0.27 vs 0.47), WavLM (0.22 vs 0.50).\n",
-    "- Catastrophically mis-evaluate XLS-R (0.03 vs 0.46) -- reporting XLS-R as worse than chance, when it is actually competitive with wav2vec2/HuBERT.\n",
-    "- Be unfair to the SSL family overall, inflating the supervised-vs-SSL gap from ~0.14 to ~0.38.\n",
-    "\n",
-    "**For the paper:** the main benchmark numbers (Sections 3-12) already use nested-CV best-layer for SSL. Section 13 is a methodological supplement that documents this choice and provides the per-layer transparency that reviewers will want. We also report last-layer r as a separate column in the grand summary table (Section 14) for anyone interested in zero-shot / out-of-the-box SSL performance."
-   ]
-  },
   {
    "cell_type": "markdown",
    "id": "part_VI_divider",
@@ -4418,42 +3003,6 @@
     "print('=' * 115)\n",
     "print(summary.round(4).to_string(index=False))\n"
    ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "cell_042",
-   "metadata": {},
-   "source": [
-    "### Key Findings\n",
-    "\n",
-    "1. **Supervised speaker embeddings predict human identity judgments best, but SSL models are competitive with fair layer selection.** On the 124,876-judgment full dataset, top supervised models (resemblyzer r=0.656, ECAPA-TDNN r=0.646) are at the boundary of the top supervised tier (BH q=0.049, marginally significant). Best-layer SSL models (WavLM r=0.510, wav2vec2 r=0.471, XLS-R r=0.465, HuBERT r=0.458) are 0.15-0.20 below the top supervised tier -- a real but substantially smaller gap than last-layer SSL comparisons would suggest. Whisper (r=0.226) is the exception: its multitask ASR training suppresses speaker identity at all layers.\n",
-    "\n",
-    "2. **Significance matters.** Of 45 pairwise model comparisons, 43 are significant after Benjamini-Hochberg FDR correction at $\\alpha=0.05$. The two non-significant pairs sit within the top supervised tier (resemblyzer/ECAPA-TDNN) or within middle-SSL clusters.\n",
-    "\n",
-    "3. **The best supervised models match or exceed the majority of individual human listeners.** resemblyzer agrees with the leave-one-out population consensus more often than 82% of individual listeners; ECAPA-TDNN beats 78%; TitaNet, RawNet3, and x-vector each beat 72-76%. State-of-the-art supervised speaker embeddings agree with population-consensus voice-identity labels more often than the majority of participants.\n",
-    "\n",
-    "4. **Best-layer SSL models fall in the middle of the human distribution.** HuBERT beats 37% of individual listeners, wav2vec2 31%, WavLM 28%, XLS-R 25%, Whisper 14%. Fair layer selection moves SSL models from \"almost no one beats them\" (last-layer protocol) to \"a third of individuals beats them.\"\n",
-    "\n",
-    "5. **The 61%-of-ceiling gap is listener variability, not representational inadequacy.** The R^2 ceiling is 0.705 (from split-half reliability); the best model sits at 0.430. Because individual humans also fall short of the population-average P(same) (mean individual agreement 0.73 vs best model 0.80), closing the gap requires modeling listener-level heterogeneity rather than a better speaker embedding.\n",
-    "\n",
-    "6. **Fair SSL evaluation requires non-default layer selection.** Using `last_hidden_state` (the HuggingFace default) drops SSL alignment by 0.16-0.43 vs best-layer selection (e.g., XLS-R: r=0.03 last layer vs r=0.46 layer 1). Section 13 documents this; the main benchmark uses nested speaker-CV best-layer selection throughout.\n",
-    "\n",
-    "7. **Calibration is mostly a scale problem.** Raw ECE is dominated by cosine-similarity-range mismatch. After Platt scaling, four of five supervised models achieve ECE_cal of 0.018-0.027, and four SSL models match (0.022-0.035). **x-vector is the lone calibration outlier** (ECE_cal = 0.136) even after Platt scaling.\n",
-    "\n",
-    "8. **Real-to-synthetic transfer is lossy but no longer catastrophic.** With best-layer SSL, R^2_transfer is near zero for all non-Whisper models (0.003-0.30); the main issue is slope/intercept shifts in the cosine-similarity-to-P(same) mapping. Whisper remains an outlier with R^2_transfer = -0.15. **Takeaway:** real-speech calibration is not a reliable proxy for synthetic-speech calibration; application-specific calibration on the relevant stimulus distribution is required.\n",
-    "\n",
-    "9. **Speaker-level population structure is captured well across model families.** Aggregating to speaker-pair RDM raises every model's alignment (best-layer SSL reaches 0.63-0.72 vs supervised 0.74-0.81). Whisper remains an outlier (0.40).\n",
-    "\n",
-    "10. **The benchmark is robust to Type 6 dominance.** Removing the 8,100 blended pairs yields rank correlation 0.976 with full-set rankings; absolute r increases for every model except Whisper, indicating Type 6 makes the benchmark harder, not easier.\n",
-    "\n",
-    "11. **Best-layer SSL largely eliminates gender disparities.** All 9 non-Whisper models have |r_M - r_F| < 0.015; Whisper alone shows a 0.12 gender disparity. Sociophonetic group disparities (~0.11-0.17) are uniform across models, suggesting stimulus-level rather than model-specific variability.\n",
-    "\n",
-    "12. **Models weakly predict human uncertainty.** Best-model confidence-vs-agreement correlation is 0.34 (resemblyzer). No model reliably flags pairs humans will find ambiguous.\n",
-    "\n",
-    "13. **Supervised embedding spaces are isotropic w.r.t. human perception; best-layer SSL is modestly anisotropic.** Diagonal Mahalanobis does not help supervised models (deltas ~-0.03 to +0.01) but yields small gains for best-layer SSL (+0.02 to +0.10). The gain does not close the supervised gap.\n",
-    "\n",
-    "14. **Voice clones are the hardest stimulus type.** Type 3 (same-speaker clones) has 44% high-disagreement pairs and shows systematically lower model-human alignment than real recordings of the same speaker."
-   ]
   }
  ],
  "metadata": {