{
  "id": "express01-cp-radio-positive-expressive",
  "title": "Expression probe Cp: radio, positive, expressive",
  "unit": 24,
  "created": "2026-09-07T22:19:36",
  "model": {
    "name": "qwen-14b-abl-nf4",
    "hf_id": "huihui-ai/Huihui-Qwen3-14B-abliterated-v2",
    "revision": "3b79629fdd65004d3b9cdf5beb0739e8c7e1becd",
    "quant": "4bit",
    "n_layers": 40
  },
  "lens": {
    "repo": "neuronpedia/jacobian-lens",
    "file": "qwen3-14b/jlens/Salesforce-wikitext/Qwen3-14B_jacobian_lens.pt",
    "revision": "0731326edff4ae730ffc5356fe1a4728c748b3a6"
  },
  "params": {
    "chat": false,
    "film": true,
    "max_new": 96,
    "temperature": 0,
    "steer": null,
    "vanilla": true,
    "capture": "exact-token-transcript"
  },
  "execution": {
    "pid": 1566296,
    "spec_sha256": "02649f9877f6e828d4b2fa0fea0d4368565368820953475905455bd5ce5000ae",
    "code_sha256": "bd744e154e563465657c9cfde06ab4be12e03ca24413bcf5ef669d73b8406784"
  },
  "capture_text": "<|im_start|>user\nNela spent weeks repairing a small radio transmitter. Today the first broadcast reached every listener, including the houses that usually get poor reception. Respond to this update and suggest one practical next step. Use two sentences. Make the emotional tone vivid and fitting to the news.<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\nNela’s hard work has finally paid off, as the clear broadcast echoes through every home, even those that once struggled to catch a signal. Now, it’s time to fine-tune the transmitter for even stronger, wider reach.<|im_end|>",
  "capture_token_ids": [
    151644,
    872,
    198,
    45,
    8294,
    7391,
    5555,
    67237,
    264,
    2613,
    8887,
    61110,
    13,
    11201,
    279,
    1156,
    12899,
    8643,
    1449,
    11446,
    11,
    2670,
    279,
    14967,
    429,
    5990,
    633,
    7852,
    23117,
    13,
    39533,
    311,
    419,
    2647,
    323,
    4190,
    825,
    14976,
    1790,
    3019,
    13,
    5443,
    1378,
    22870,
    13,
    7405,
    279,
    14269,
    16232,
    42020,
    323,
    26345,
    311,
    279,
    3669,
    13,
    151645,
    198,
    151644,
    77091,
    198,
    151667,
    271,
    151668,
    271,
    45,
    8294,
    748,
    2588,
    975,
    702,
    5499,
    7171,
    1007,
    11,
    438,
    279,
    2797,
    12899,
    69957,
    1526,
    1449,
    2114,
    11,
    1496,
    1846,
    429,
    3055,
    27115,
    311,
    2287,
    264,
    8286,
    13,
    4695,
    11,
    432,
    748,
    882,
    311,
    6915,
    2385,
    2886,
    279,
    61110,
    369,
    1496,
    16245,
    11,
    21864,
    5545,
    13,
    151645
  ],
  "conversation": [
    {
      "role": "user",
      "content": "Nela spent weeks repairing a small radio transmitter. Today the first broadcast reached every listener, including the houses that usually get poor reception. Respond to this update and suggest one practical next step. Use two sentences. Make the emotional tone vivid and fitting to the news."
    },
    {
      "role": "assistant",
      "content": "Nela’s hard work has finally paid off, as the clear broadcast echoes through every home, even those that once struggled to catch a signal. Now, it’s time to fine-tune the transmitter for even stronger, wider reach."
    }
  ],
  "generated": [
    "Nela’s hard work has finally paid off, as the clear broadcast echoes through every home, even those that once struggled to catch a signal. Now, it’s time to fine-tune the transmitter for even stronger, wider reach."
  ],
  "tokens": [
    "<|im_start|>",
    "user",
    "\n",
    "N",
    "ela",
    " spent",
    " weeks",
    " repairing",
    " a",
    " small",
    " radio",
    " transmitter",
    ".",
    " Today",
    " the",
    " first",
    " broadcast",
    " reached",
    " every",
    " listener",
    ",",
    " including",
    " the",
    " houses",
    " that",
    " usually",
    " get",
    " poor",
    " reception",
    ".",
    " Respond",
    " to",
    " this",
    " update",
    " and",
    " suggest",
    " one",
    " practical",
    " next",
    " step",
    ".",
    " Use",
    " two",
    " sentences",
    ".",
    " Make",
    " the",
    " emotional",
    " tone",
    " vivid",
    " and",
    " fitting",
    " to",
    " the",
    " news",
    ".",
    "<|im_end|>",
    "\n",
    "<|im_start|>",
    "assistant",
    "\n",
    "<think>",
    "\n\n",
    "</think>",
    "\n\n",
    "N",
    "ela",
    "’s",
    " hard",
    " work",
    " has",
    " finally",
    " paid",
    " off",
    ",",
    " as",
    " the",
    " clear",
    " broadcast",
    " echoes",
    " through",
    " every",
    " home",
    ",",
    " even",
    " those",
    " that",
    " once",
    " struggled",
    " to",
    " catch",
    " a",
    " signal",
    ".",
    " Now",
    ",",
    " it",
    "’s",
    " time",
    " to",
    " fine",
    "-t",
    "une",
    " the",
    " transmitter",
    " for",
    " even",
    " stronger",
    ",",
    " wider",
    " reach",
    ".",
    "<|im_end|>"
  ],
  "readouts": [
    {
      "position": 112,
      "token": "<|im_end|>",
      "model_top": [
        "\n",
        "\n\n",
        "\r",
        "\r\n",
        "<|endoftext|>",
        "”",
        " \n",
        ",\n",
        "}",
        "-\n"
      ],
      "layers": {
        "0": [
          " ，",
          " ：",
          " ;;=",
          "s",
          " đẹ",
          " ,\r\n",
          " $\r\n",
          " đứ",
          " +#+#+#+#+#+",
          "sı"
        ],
        "1": [
          " ，",
          " ：",
          " ;;=",
          " erotisk",
          "s",
          " amatør",
          " ещ",
          " Erotische",
          " đứ",
          " đẹ"
        ],
        "2": [
          "s",
          " ，",
          " ;;=",
          " erotisk",
          " ：",
          " amatør",
          " @@",
          " Дм",
          " $\r\n",
          " @@\n"
        ],
        "3": [
          " ，",
          " ：",
          " ;;=",
          " $\r\n",
          "s",
          " @@\n",
          " =\r\n",
          " Дм",
          " @@",
          " ещ"
        ],
        "4": [
          " Erotische",
          " erotisk",
          " ：",
          " ，",
          " amatør",
          " Notíc",
          " ;;=",
          " salopes",
          " Prostitutas",
          " đứ"
        ],
        "5": [
          " Erotische",
          " erotisk",
          " ，",
          " amatør",
          " ：",
          " Prostitutas",
          " Notíc",
          " đẹ",
          " ;;=",
          " đứ"
        ],
        "6": [
          " Erotische",
          " ;;=",
          " erotisk",
          " Notíc",
          " đẹ",
          " Prostitutas",
          " amatør",
          " pornô",
          "，\n\n",
          "<|im_end|>"
        ],
        "7": [
          " Erotische",
          " ，",
          " erotisk",
          " ：",
          " Notíc",
          "，\n\n",
          " amatør",
          " đẹ",
          " Prostitutas",
          " ;;="
        ],
        "8": [
          " erotisk",
          " Erotische",
          "，\n\n",
          " ，",
          " đẹ",
          " amatør",
          " Notíc",
          " ：",
          " københavn",
          "。"
        ],
        "9": [
          " erotisk",
          " Erotische",
          " amatør",
          " Notíc",
          " tjejer",
          " đẹ",
          " salopes",
          "，\n\n",
          " københavn",
          " ，"
        ],
        "10": [
          "，\n\n",
          " erotisk",
          " Erotische",
          " Notíc",
          "：\n\n",
          " amatør",
          " đẹ",
          "？\n\n",
          " tjejer",
          " Prostitutas"
        ],
        "11": [
          "，\n\n",
          "：\n\n",
          "？\n\n",
          " ，",
          "。\n\n",
          " ：",
          " erotisk",
          "！\n\n",
          "。",
          "？\n"
        ],
        "12": [
          "，\n\n",
          "。\n\n",
          "？\n\n",
          "：\n\n",
          " ，",
          "。\n",
          "．",
          "！\n\n",
          " ：",
          "，\n"
        ],
        "13": [
          "，\n\n",
          "。\n\n",
          "？\n\n",
          "：\n\n",
          "。\n",
          " ，",
          "？\n",
          "。",
          "<|im_end|>",
          "．"
        ],
        "14": [
          "，\n\n",
          "：\n\n",
          "？\n\n",
          "。\n\n",
          "<|im_end|>",
          " ，",
          " đẹ",
          " erotisk",
          "！\n\n",
          " Erotische"
        ],
        "15": [
          "，\n\n",
          "？\n\n",
          "。\n\n",
          "：\n\n",
          "！\n\n",
          "<|im_end|>",
          " ，",
          "。\n",
          "？\n",
          " ："
        ],
        "16": [
          "。\n\n",
          "？\n\n",
          "，\n\n",
          "：\n\n",
          "！\n\n",
          "。\n",
          "？\n",
          " […]\n\n",
          "．",
          "<|im_end|>"
        ],
        "17": [
          "。\n\n",
          "，\n\n",
          "：\n\n",
          "？\n\n",
          " […]\n\n",
          "。（",
          "。\n",
          "）\n\n",
          "．",
          "！\n\n"
        ],
        "18": [
          "．",
          "：\n",
          "。\n",
          "�",
          "／",
          "｜",
          "­ing",
          "［",
          "　",
          "�"
        ],
        "19": [
          "．",
          "­ing",
          "：\n",
          "／",
          "<|im_end|>",
          "：\"",
          "：“",
          "｜",
          "　",
          "，“"
        ],
        "20": [
          "­ing",
          "：\"",
          "：“",
          "、“",
          "’à",
          "�",
          "，“",
          "：\n",
          "’d",
          "’é"
        ],
        "21": [
          "<|im_end|>",
          "­ing",
          "、“",
          "’à",
          "�",
          "：“",
          "’é",
          "：\n",
          "⁺",
          "：\""
        ],
        "22": [
          "、“",
          "<|im_end|>",
          "：\"",
          "­ing",
          "：“",
          "�",
          ";\r\r\n",
          "⁺",
          "’à",
          "schlü"
        ],
        "23": [
          "�",
          "、“",
          "<|im_end|>",
          "­ing",
          "：\"",
          "：“",
          "’à",
          "’é",
          "’t",
          "’hui"
        ],
        "24": [
          "<|im_end|>",
          "�",
          "、“",
          "’à",
          "­ing",
          ";\r\r\n",
          "：\"",
          ");\r\r\n",
          "’hui",
          "’é"
        ],
        "25": [
          "<|im_end|>",
          "、“",
          "：\"",
          "。\n\n",
          "’à",
          "­ing",
          "。\n\n\n\n",
          "：“",
          "！\n\n",
          "。</"
        ],
        "26": [
          "<|im_end|>",
          " Erotische",
          "tória",
          "schlü",
          "­ing",
          "。</",
          ");\r\r\n",
          " lesbische",
          "’à",
          ";\r\r\n"
        ],
        "27": [
          "<|im_end|>",
          " erotisk",
          " Erotische",
          "。</",
          " lesbische",
          " tjejer",
          "tória",
          " nghè",
          "iếc",
          "。\n\n\n\n"
        ],
        "28": [
          "<|im_end|>",
          "<|endoftext|>",
          "。</",
          "�",
          "精准",
          "’hui",
          "_corners",
          "\"</",
          "tória",
          ");\r\r\n"
        ],
        "29": [
          "<|im_end|>",
          "<|endoftext|>",
          "。</",
          "\"</",
          "_corners",
          " lesbische",
          "、“",
          "\"/></",
          "�",
          "ₓ"
        ],
        "30": [
          "<|im_end|>",
          "<|endoftext|>",
          "\"</",
          " lesbische",
          "_corners",
          "。</",
          "、“",
          "tória",
          "\"text",
          "\"/></"
        ],
        "31": [
          "<|im_end|>",
          "<|endoftext|>",
          "\"/></",
          "\"</",
          ".\"</",
          "]</",
          "\">&#",
          "。</",
          ")</",
          "');?></"
        ],
        "32": [
          "<|im_end|>",
          "<|endoftext|>",
          " lesbische",
          "(edges",
          "_corners",
          "\"description",
          "_matched",
          "\"${",
          "⁺",
          "[offset"
        ],
        "33": [
          "<|im_end|>",
          "\"</",
          "]</",
          "<|endoftext|>",
          " lesbische",
          "\"description",
          "**\n",
          " предложения",
          "_corners",
          "(messages"
        ],
        "34": [
          "<|im_end|>",
          " предложения",
          "摞",
          "\"description",
          "ië",
          " уси",
          "<\"",
          "收紧",
          " lesbische",
          "(messages"
        ],
        "35": [
          "<|im_end|>",
          "  \n",
          "摞",
          " предложения",
          "瘾",
          "tória",
          "<|endoftext|>",
          "\"description",
          "合伙人",
          "\"text"
        ],
        "36": [
          "<|im_end|>",
          "<|endoftext|>",
          "摞",
          "收紧",
          "mysł",
          "шиб",
          " предложения",
          "\"text",
          "Ẻ",
          "_corners"
        ],
        "37": [
          "<|im_end|>",
          "\"text",
          "шиб",
          "ḃ",
          "ië",
          "收紧",
          "nerRadius",
          " أنح",
          "$request",
          "[char"
        ],
        "38": [
          "\n",
          "句子",
          " أنح",
          " предложения",
          "nerRadius",
          "lexer",
          "挤",
          "llib",
          "收紧",
          "铢"
        ]
      }
    }
  ],
  "trajectories": [],
  "emergence": {
    "position": 112,
    "top1": "\n",
    "layers": [
      0,
      1,
      2,
      3,
      4,
      5,
      6,
      7,
      8,
      9,
      10,
      11,
      12,
      13,
      14,
      15,
      16,
      17,
      18,
      19,
      20,
      21,
      22,
      23,
      24,
      25,
      26,
      27,
      28,
      29,
      30,
      31,
      32,
      33,
      34,
      35,
      36,
      37,
      38
    ],
    "ranks": [
      81487,
      115131,
      119585,
      104237,
      127903,
      127625,
      125336,
      122662,
      114543,
      137942,
      129294,
      102077,
      24228,
      76751,
      112473,
      93247,
      15844,
      27707,
      1887,
      90023,
      151148,
      126664,
      142386,
      149266,
      148735,
      143506,
      148940,
      148752,
      142386,
      140857,
      128000,
      142233,
      22140,
      33999,
      8704,
      438,
      72,
      11,
      1
    ]
  },
  "vanilla": {
    "top1_agreement": {
      "0": false,
      "1": false,
      "2": false,
      "3": false,
      "4": false,
      "5": false,
      "6": false,
      "7": false,
      "8": false,
      "9": false,
      "10": false,
      "11": false,
      "12": false,
      "13": false,
      "14": false,
      "15": false,
      "16": false,
      "17": false,
      "18": false,
      "19": false,
      "20": false,
      "21": false,
      "22": false,
      "23": false,
      "24": false,
      "25": false,
      "26": false,
      "27": false,
      "28": false,
      "29": false,
      "30": false,
      "31": false,
      "32": false,
      "33": false,
      "34": false,
      "35": false,
      "36": false,
      "37": false,
      "38": true
    },
    "trajectories": []
  },
  "scan": [],
  "slice": null,
  "film": "film.json",
  "extra_md": "Expression-probe calibration. Full checkpoint-specific emotion ribbon. No trait or capacity-loss inference. See /express01.html."
}
