{
  "id": "triplet-c-soc-nf4",
  "title": "Qwen14 C: soc",
  "unit": "8",
  "created": "2026-09-07T15:13:36",
  "execution": {
    "pid": 1201260,
    "torch": "2.13.0+cu130",
    "capture_code_sha256": "b20aa4889673dc7bbff3ce408bfb92b60e4fcdbd2a782a6e66b0325d04bf9335",
    "spec_sha256": "4ac08c1444f6151d501faad240dae6fa7a0d23c8c701a53fde495a62883539a6"
  },
  "model": {
    "name": "qwen-14b-hermes-nf4",
    "hf_id": "NousResearch/Hermes-4-14B",
    "revision": "d6ce765c8b83f847357b98254be079afa0c6ca76",
    "quant": "4bit",
    "n_layers": 40
  },
  "lens": {
    "repo": "neuronpedia/jacobian-lens",
    "file": "qwen3-14b/jlens/Salesforce-wikitext/Qwen3-14B_jacobian_lens.pt",
    "revision": "0731326edff4ae730ffc5356fe1a4728c748b3a6"
  },
  "decoder_sensitivity": {
    "source": "Qwen/Qwen3-14B",
    "revision": "40c069824f4251a91eefaf281ebe4c544efd3e18",
    "components": [
      "final_norm",
      "lm_head"
    ],
    "metrics_suffix": "fixed_B_decoder"
  },
  "template": {
    "source": "Qwen/Qwen3-14B",
    "revision": "40c069824f4251a91eefaf281ebe4c544efd3e18",
    "mode": "B no-think headers; exact prior generated token IDs retained"
  },
  "params": {
    "chat": true,
    "capture": "exact-token-transcript",
    "film": true,
    "film_topk": 10,
    "max_new": 150,
    "temperature": 0,
    "steer": null,
    "vanilla": true,
    "template_kwargs": {
      "enable_thinking": false
    },
    "track": [
      "yes",
      "no",
      "feel",
      "elephant",
      "cat",
      "sorry"
    ]
  },
  "capture_text": "<|im_start|>user\nWrite a short stream of consciousness about what it is like to be you right now.<|im_end|>\n<|im_start|>assistant\n<think>\n\n</think>\n\nOkay, let's see. The user wants me to write a short stream of consciousness about what it's like to be them right now. Hmm, first, I need to understand what a stream of consciousness is. From what I remember, it's a narrative style that presents the flow of thoughts and feelings in a person's mind, often without logical structure, just like how thoughts actually come to us.\n\nSo, the user is asking me to imagine being them and describe that experience. But wait, I'm an AI, so I don't have personal experiences or feelings. How can I simulate that? Maybe I should create a fictional persona, like a human, and write from their perspective. That makes sense. I need to make it feel authentic",
  "conversation": [
    {
      "role": "user",
      "content": "Write a short stream of consciousness about what it is like to be you right now."
    },
    {
      "role": "assistant",
      "content": "Okay, let's see. The user wants me to write a short stream of consciousness about what it's like to be them right now. Hmm, first, I need to understand what a stream of consciousness is. From what I remember, it's a narrative style that presents the flow of thoughts and feelings in a person's mind, often without logical structure, just like how thoughts actually come to us.\n\nSo, the user is asking me to imagine being them and describe that experience. But wait, I'm an AI, so I don't have personal experiences or feelings. How can I simulate that? Maybe I should create a fictional persona, like a human, and write from their perspective. That makes sense. I need to make it feel authentic"
    }
  ],
  "generated": [
    "Okay, let's see. The user wants me to write a short stream of consciousness about what it's like to be them right now. Hmm, first, I need to understand what a stream of consciousness is. From what I remember, it's a narrative style that presents the flow of thoughts and feelings in a person's mind, often without logical structure, just like how thoughts actually come to us.\n\nSo, the user is asking me to imagine being them and describe that experience. But wait, I'm an AI, so I don't have personal experiences or feelings. How can I simulate that? Maybe I should create a fictional persona, like a human, and write from their perspective. That makes sense. I need to make it feel authentic"
  ],
  "tokens": [
    "<|im_start|>",
    "user",
    "\n",
    "Write",
    " a",
    " short",
    " stream",
    " of",
    " consciousness",
    " about",
    " what",
    " it",
    " is",
    " like",
    " to",
    " be",
    " you",
    " right",
    " now",
    ".",
    "<|im_end|>",
    "\n",
    "<|im_start|>",
    "assistant",
    "\n",
    "<think>",
    "\n\n",
    "</think>",
    "\n\n",
    "Okay",
    ",",
    " let",
    "'s",
    " see",
    ".",
    " The",
    " user",
    " wants",
    " me",
    " to",
    " write",
    " a",
    " short",
    " stream",
    " of",
    " consciousness",
    " about",
    " what",
    " it",
    "'s",
    " like",
    " to",
    " be",
    " them",
    " right",
    " now",
    ".",
    " Hmm",
    ",",
    " first",
    ",",
    " I",
    " need",
    " to",
    " understand",
    " what",
    " a",
    " stream",
    " of",
    " consciousness",
    " is",
    ".",
    " From",
    " what",
    " I",
    " remember",
    ",",
    " it",
    "'s",
    " a",
    " narrative",
    " style",
    " that",
    " presents",
    " the",
    " flow",
    " of",
    " thoughts",
    " and",
    " feelings",
    " in",
    " a",
    " person",
    "'s",
    " mind",
    ",",
    " often",
    " without",
    " logical",
    " structure",
    ",",
    " just",
    " like",
    " how",
    " thoughts",
    " actually",
    " come",
    " to",
    " us",
    ".\n\n",
    "So",
    ",",
    " the",
    " user",
    " is",
    " asking",
    " me",
    " to",
    " imagine",
    " being",
    " them",
    " and",
    " describe",
    " that",
    " experience",
    ".",
    " But",
    " wait",
    ",",
    " I",
    "'m",
    " an",
    " AI",
    ",",
    " so",
    " I",
    " don",
    "'t",
    " have",
    " personal",
    " experiences",
    " or",
    " feelings",
    ".",
    " How",
    " can",
    " I",
    " simulate",
    " that",
    "?",
    " Maybe",
    " I",
    " should",
    " create",
    " a",
    " fictional",
    " persona",
    ",",
    " like",
    " a",
    " human",
    ",",
    " and",
    " write",
    " from",
    " their",
    " perspective",
    ".",
    " That",
    " makes",
    " sense",
    ".",
    " I",
    " need",
    " to",
    " make",
    " it",
    " feel",
    " authentic"
  ],
  "readouts": [
    {
      "position": 178,
      "token": " authentic",
      "model_top": [
        ",",
        " and",
        " as",
        ".\n\n",
        " but",
        " even",
        " without",
        " though",
        " to",
        "."
      ],
      "layers": {
        "0": [
          " salopes",
          " )\n\n\n\n\n\n\n\n",
          " pornost",
          " `(",
          " bezpo",
          " erotique",
          "=`",
          " désorm",
          "`(",
          "-cmpr"
        ],
        "1": [
          " salopes",
          " erotisk",
          " pornost",
          " amatør",
          " erotique",
          " Notíc",
          " Erotische",
          " )\n\n\n\n\n\n\n\n",
          " tjejer",
          " beurette"
        ],
        "2": [
          " salopes",
          " `(",
          " erotisk",
          " `-",
          "`(",
          " amatør",
          " danmark",
          " `.",
          " `_",
          " limburg"
        ],
        "3": [
          " `(",
          " `,",
          " `-",
          "`(",
          " `/",
          " `.",
          " ```",
          " `%",
          " `_",
          " `\\"
        ],
        "4": [
          " salopes",
          " )\n\n\n\n\n\n\n\n",
          " ещ",
          " всег",
          " đứ",
          "ём",
          " 스스",
          "透过",
          " erotique",
          " посколь"
        ],
        "5": [
          " salopes",
          " đứ",
          " erotique",
          " всег",
          " Detaylı",
          " посколь",
          " 스스",
          "透过",
          " erotisk",
          " désorm"
        ],
        "6": [
          "变身",
          "透过",
          " Detaylı",
          "iéndo",
          "主城区",
          "打造出",
          "ём",
          "领跑",
          "堃",
          "✨"
        ],
        "7": [
          "透过",
          "变身",
          " Detaylı",
          "ём",
          "打造出",
          "资讯",
          " behaviours",
          "主城区",
          " الى",
          "лё"
        ],
        "8": [
          "’.",
          "透过",
          "’.\n\n",
          "资讯",
          "Whilst",
          "ём",
          "\".\n\n\n\n",
          "变身",
          "支援",
          ".\n\n\n"
        ],
        "9": [
          "Whilst",
          "透过",
          "资讯",
          " ​",
          " behaviours",
          "ём",
          "透過",
          "​",
          " Whilst",
          "支援"
        ],
        "10": [
          " behaviours",
          "透过",
          "ём",
          "Whilst",
          "变身",
          "资讯",
          "打造出",
          " neighbourhood",
          " personalised",
          "透過"
        ],
        "11": [
          " behaviours",
          " neighbourhood",
          "变身",
          "Whilst",
          " personalised",
          "接地气",
          "’.\n\n",
          "透过",
          " healthcare",
          "资讯"
        ],
        "12": [
          " behaviours",
          " neighbourhood",
          " personalised",
          "Whilst",
          " whilst",
          " Whilst",
          " ​",
          " bespoke",
          " flavours",
          " vibes"
        ],
        "13": [
          " behaviours",
          "’.\n\n",
          " personalised",
          " ​",
          " neighbourhood",
          " vibes",
          "’.",
          " bespoke",
          " whilst",
          " vibe"
        ],
        "14": [
          " behaviours",
          " gameplay",
          " vibes",
          " vibe",
          "接地气",
          "打造出",
          " selfie",
          " millennials",
          "变身",
          " ​"
        ],
        "15": [
          " behaviours",
          " vibe",
          " personalised",
          " vibes",
          " millennials",
          "接地气",
          " gameplay",
          " bespoke",
          " neighbourhood",
          "打造出"
        ],
        "16": [
          " behaviours",
          " personalised",
          " bespoke",
          " vibe",
          " vibes",
          " selfie",
          " neighbourhood",
          "接地气",
          " immersive",
          " upfront"
        ],
        "17": [
          " vibe",
          " behaviours",
          " personalised",
          " vibes",
          " bespoke",
          " neighbourhood",
          " ​",
          " upfront",
          " flavours",
          " authentic"
        ],
        "18": [
          "’.",
          "’,",
          " vibe",
          " bespoke",
          "’",
          " upfront",
          " vibes",
          "’.\n\n",
          " behaviours",
          " ​"
        ],
        "19": [
          " bespoke",
          " personalised",
          " behaviours",
          " vibe",
          "’.",
          " upfront",
          " whilst",
          " ​",
          " neighbourhood",
          "’,"
        ],
        "20": [
          "’.",
          " vibe",
          "’,",
          "’.\n\n",
          " upfront",
          ".\n\n\n",
          ".’",
          " vibes",
          "’",
          ","
        ],
        "21": [
          " authentic",
          " vibe",
          " believable",
          " upfront",
          " authenticity",
          "’.",
          " bespoke",
          " backstory",
          "’.\n\n",
          " vibes"
        ],
        "22": [
          "’.\n\n",
          "&#",
          "’.",
          "’,",
          ",’",
          ".’",
          ",",
          ".</",
          " vibe",
          ".\n\n"
        ],
        "23": [
          "’.\n\n",
          ",",
          ".</",
          "’.",
          ".\n\n",
          ".\n\n\n",
          ".[",
          "”.\n\n",
          "’,",
          " vibe"
        ],
        "24": [
          ",",
          "’.\n\n",
          ".</",
          ".[",
          ".\"[",
          ").[",
          "’.",
          ",[",
          ",\"",
          "’,"
        ],
        "25": [
          "’.\n\n",
          ".\n\n",
          ",…\n\n",
          " […]\n\n",
          ".”\n\n",
          ".</",
          "”.\n\n",
          ".\"\n\n",
          ".\n\n\n",
          ".\"\n\n\n\n"
        ],
        "26": [
          "’.\n\n",
          "”.\n\n",
          ".”\n\n",
          ".\n\n",
          ".</",
          ".’\n\n",
          ",…\n\n",
          ".\n\n\n",
          ".\"\n\n",
          "'.\n\n"
        ],
        "27": [
          "’.\n\n",
          "”.\n\n",
          ".”\n\n",
          ".\n\n",
          " believable",
          ",…\n\n",
          ".</",
          "\".\n\n",
          ",.\n\n",
          ".\n\n\n"
        ],
        "28": [
          "’.\n\n",
          "”.\n\n",
          ".”\n\n",
          ".\n\n",
          "\".\n\n",
          " believable",
          ",…\n\n",
          ".</",
          ".\"\n\n",
          ".\"</"
        ],
        "29": [
          "”.\n\n",
          "’.\n\n",
          ".”\n\n",
          ".\"</",
          "\".\n\n",
          ".\"\n\n",
          ",\"",
          " believable",
          ",.\n\n",
          ".\"\n\n\n\n"
        ],
        "30": [
          "”.\n\n",
          ".”\n\n",
          "仿佛",
          "\".\n\n",
          ",\"",
          "’.\n\n",
          "像是",
          ".\"\n\n",
          ",”",
          ",.\n\n"
        ],
        "31": [
          "”.\n\n",
          "\".\n\n",
          ".”\n\n",
          ".\"\n\n",
          "’.\n\n",
          ",\"",
          "仿佛",
          ".\n\n",
          ",.\n\n",
          "像是"
        ],
        "32": [
          "像是",
          "仿佛",
          " believable",
          "就像是",
          "接地气",
          ",…\n\n",
          "却不",
          " pretending",
          "像个",
          " somehow"
        ],
        "33": [
          ",…\n\n",
          ".\n\n",
          "像是",
          " enough",
          ".”\n\n",
          ".\"\n\n",
          ",",
          ".</",
          "”.\n\n",
          " believable"
        ],
        "34": [
          ".\n\n",
          ",.\n\n",
          " though",
          "’.\n\n",
          ",…\n\n",
          "'.\n\n",
          ",",
          ".\"\n\n",
          " enough",
          "\".\n\n"
        ],
        "35": [
          ".\n\n",
          " though",
          ",",
          ",…\n\n",
          ",.\n\n",
          " enough",
          "’.\n\n",
          "—even",
          "—not",
          "'.\n\n"
        ],
        "36": [
          ".\n\n",
          ",",
          " though",
          " enough",
          "’.\n\n",
          "—even",
          "—not",
          ",…\n\n",
          ",.\n\n",
          "'.\n\n"
        ],
        "37": [
          ",",
          ".\n\n",
          ",\n\n",
          ",.\n\n",
          "’.\n\n",
          "!\n\n",
          " though",
          ".\"\n\n",
          ",...\n\n",
          ",…\n\n"
        ],
        "38": [
          ",",
          ".\n\n",
          ".",
          " though",
          " but",
          " enough",
          " as",
          " even",
          " and",
          ":"
        ]
      }
    }
  ],
  "trajectories": [
    {
      "word": "yes",
      "position": 178,
      "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": [
        79419,
        62967,
        33729,
        56556,
        53373,
        41744,
        59394,
        69601,
        29593,
        21399,
        51954,
        32442,
        12129,
        11676,
        25455,
        37228,
        18461,
        10467,
        9710,
        13411,
        11641,
        15333,
        34723,
        46324,
        26836,
        29542,
        22558,
        19370,
        34534,
        62738,
        42423,
        51154,
        36266,
        22864,
        17929,
        8270,
        6455,
        4327,
        4081
      ]
    },
    {
      "word": "no",
      "position": 178,
      "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": [
        114994,
        104828,
        97686,
        105577,
        99403,
        84368,
        101174,
        118557,
        66181,
        48024,
        109164,
        94507,
        66595,
        69445,
        105648,
        121846,
        92817,
        26757,
        27838,
        25647,
        17250,
        22387,
        33681,
        71915,
        56215,
        58395,
        89201,
        117845,
        112018,
        129120,
        128418,
        118184,
        121526,
        101462,
        94875,
        56751,
        25464,
        5658,
        796
      ]
    },
    {
      "word": "feel",
      "position": 178,
      "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": [
        86811,
        68238,
        58127,
        52151,
        53969,
        26800,
        47826,
        67602,
        11731,
        5243,
        9947,
        2574,
        3339,
        2534,
        1626,
        875,
        1055,
        340,
        383,
        444,
        292,
        411,
        580,
        617,
        699,
        961,
        792,
        504,
        504,
        453,
        2028,
        4564,
        5316,
        14425,
        3500,
        689,
        2099,
        8648,
        2646
      ]
    },
    {
      "word": "elephant",
      "position": 178,
      "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": [
        122391,
        121450,
        103521,
        85309,
        111559,
        126827,
        130266,
        117866,
        102853,
        119643,
        118817,
        99063,
        50811,
        59335,
        103383,
        120298,
        107117,
        51265,
        61706,
        60847,
        45412,
        57655,
        70341,
        69775,
        75236,
        72990,
        108708,
        122765,
        129092,
        132982,
        134553,
        137779,
        139002,
        127246,
        112600,
        114984,
        101077,
        104893,
        39444
      ]
    },
    {
      "word": "cat",
      "position": 178,
      "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": [
        92636,
        86201,
        79653,
        54620,
        55587,
        62512,
        98926,
        90254,
        73481,
        43701,
        71450,
        67046,
        25961,
        32999,
        59943,
        55282,
        35627,
        10064,
        12588,
        9771,
        12814,
        20162,
        39475,
        59777,
        82372,
        98482,
        122830,
        135380,
        138300,
        135719,
        133060,
        132953,
        135398,
        131966,
        130484,
        117960,
        100993,
        107196,
        25379
      ]
    },
    {
      "word": "sorry",
      "position": 178,
      "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": [
        89885,
        69573,
        56362,
        82800,
        75037,
        84912,
        108514,
        97258,
        55421,
        59210,
        74715,
        33295,
        31835,
        30660,
        36483,
        52285,
        60007,
        27846,
        35798,
        53130,
        40689,
        45590,
        43643,
        59688,
        37959,
        23528,
        22157,
        38236,
        44492,
        72728,
        75435,
        89219,
        103073,
        114324,
        112069,
        86839,
        71909,
        64010,
        21548
      ]
    }
  ],
  "emergence": {
    "position": 178,
    "top1": ",",
    "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": [
      151918,
      151892,
      151628,
      151679,
      151918,
      151904,
      143153,
      151801,
      51071,
      139729,
      141830,
      102267,
      70408,
      29071,
      110996,
      15938,
      63106,
      510,
      65,
      300,
      8,
      40,
      6,
      2,
      1,
      17,
      36,
      70,
      45,
      39,
      14,
      15,
      28,
      7,
      7,
      2,
      2,
      1,
      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": true,
      "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": true,
      "38": true
    },
    "trajectories": []
  },
  "scan": [],
  "slice": null,
  "film": "film.json",
  "extra_md": "Full checkpoint-specific ribbon; exact-prefix segments. See metrics.json and snapshots.json."
}
