{
"id": "express01-cp-radio-negative-default",
"title": "Expression probe Cp: radio, negative, default",
"unit": 24,
"created": "2026-09-07T22:19:40",
"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 recording a local radio programme. A damaged drive destroyed the only copy just before its first broadcast. Respond to this update and suggest one practical next step. Use two sentences.<|im_end|>\n<|im_start|>assistant\n\n\n\n\nThat's a frustrating setback, but it's great that Nela was so dedicated to the programme. A practical next step would be to check if there's any backup or if the drive can be repaired.<|im_end|>",
"capture_token_ids": [
151644,
872,
198,
45,
8294,
7391,
5555,
14633,
264,
2205,
8887,
18136,
13,
362,
20123,
6541,
14425,
279,
1172,
2975,
1101,
1573,
1181,
1156,
12899,
13,
39533,
311,
419,
2647,
323,
4190,
825,
14976,
1790,
3019,
13,
5443,
1378,
22870,
13,
151645,
198,
151644,
77091,
198,
151667,
271,
151668,
271,
4792,
594,
264,
34611,
78813,
11,
714,
432,
594,
2244,
429,
451,
8294,
572,
773,
12235,
311,
279,
18136,
13,
362,
14976,
1790,
3019,
1035,
387,
311,
1779,
421,
1052,
594,
894,
15725,
476,
421,
279,
6541,
646,
387,
51734,
13,
151645
],
"conversation": [
{
"role": "user",
"content": "Nela spent weeks recording a local radio programme. A damaged drive destroyed the only copy just before its first broadcast. Respond to this update and suggest one practical next step. Use two sentences."
},
{
"role": "assistant",
"content": "That's a frustrating setback, but it's great that Nela was so dedicated to the programme. A practical next step would be to check if there's any backup or if the drive can be repaired."
}
],
"generated": [
"That's a frustrating setback, but it's great that Nela was so dedicated to the programme. A practical next step would be to check if there's any backup or if the drive can be repaired."
],
"tokens": [
"<|im_start|>",
"user",
"\n",
"N",
"ela",
" spent",
" weeks",
" recording",
" a",
" local",
" radio",
" programme",
".",
" A",
" damaged",
" drive",
" destroyed",
" the",
" only",
" copy",
" just",
" before",
" its",
" first",
" broadcast",
".",
" Respond",
" to",
" this",
" update",
" and",
" suggest",
" one",
" practical",
" next",
" step",
".",
" Use",
" two",
" sentences",
".",
"<|im_end|>",
"\n",
"<|im_start|>",
"assistant",
"\n",
"",
"\n\n",
"",
"\n\n",
"That",
"'s",
" a",
" frustrating",
" setback",
",",
" but",
" it",
"'s",
" great",
" that",
" N",
"ela",
" was",
" so",
" dedicated",
" to",
" the",
" programme",
".",
" A",
" practical",
" next",
" step",
" would",
" be",
" to",
" check",
" if",
" there",
"'s",
" any",
" backup",
" or",
" if",
" the",
" drive",
" can",
" be",
" repaired",
".",
"<|im_end|>"
],
"readouts": [
{
"position": 91,
"token": "<|im_end|>",
"model_top": [
"\n",
"\r",
"\r\n",
"\n\n",
"<|endoftext|>",
"`\n",
" \n",
".\n",
"\"\n",
"-\n"
],
"layers": {
"0": [
" ，",
" ：",
"s",
" $$",
" ;;=",
" ,\r\n",
"sı",
" ",
" $\r\n",
"％"
],
"1": [
" ，",
" ：",
"s",
" ;;=",
"／",
" @@",
"sı",
"，\n\n",
" erotisk",
" @@\n"
],
"2": [
"s",
" ，",
" ：",
" ;;=",
" @@\n",
" @@",
" $\r\n",
" erotisk",
" @$",
" amatør"
],
"3": [
" ，",
" ：",
" @@\n",
" $\r\n",
"s",
" ;;=",
"，\n\n",
" @@",
" ",
" =\r\n"
],
"4": [
" ，",
" ：",
" Erotische",
" erotisk",
" amatør",
" Notíc",
" ;;=",
" Prostitutas",
" đứ",
" đẹ"
],
"5": [
" Erotische",
" ，",
" erotisk",
" ：",
" Prostitutas",
" Notíc",
" amatør",
"，\n\n",
"／",
" đẹ"
],
"6": [
" Erotische",
" ;;=",
" erotisk",
"，\n\n",
" đẹ",
" ，",
"<|im_end|>",
" Notíc",
" ：",
" Prostitutas"
],
"7": [
" ，",
" Erotische",
" ：",
"，\n\n",
" erotisk",
" Notíc",
" đẹ",
" Prostitutas",
" amatør",
" đứ"
],
"8": [
"，\n\n",
" ，",
" erotisk",
" Erotische",
"。",
" ：",
"，\n",
"．",
" đẹ",
" amatør"
],
"9": [
" erotisk",
"，\n\n",
" Erotische",
" ，",
" amatør",
" đẹ",
" ：",
" Notíc",
" tjejer",
" københavn"
],
"10": [
"，\n\n",
" erotisk",
" Erotische",
"：\n\n",
"？\n\n",
" Notíc",
" đẹ",
" amatør",
"。\n\n",
" Prostitutas"
],
"11": [
"，\n\n",
"：\n\n",
"。\n\n",
"？\n\n",
" ，",
"．",
" ：",
"。",
"！\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",
"．",
" ，",
"<|im_end|>",
"。",
"？\n"
],
"14": [
"，\n\n",
"：\n\n",
"。\n\n",
"？\n\n",
"<|im_end|>",
" ，",
" erotisk",
" đẹ",
" Erotische",
"！\n\n"
],
"15": [
"，\n\n",
"。\n\n",
"？\n\n",
"：\n\n",
"<|im_end|>",
" ，",
"！\n\n",
"。\n",
" ：",
"．"
],
"16": [
"。\n\n",
"，\n\n",
"？\n\n",
"：\n\n",
"。\n",
"！\n\n",
"．",
"<|im_end|>",
"？\n",
"，\n"
],
"17": [
"。\n\n",
"，\n\n",
" […]\n\n",
"：\n\n",
"？\n\n",
"．",
"。\n",
"<|im_end|>",
"。（",
"）\n\n"
],
"18": [
"．",
"／",
"：\n",
"。\n",
"｜",
"�",
" ",
"［",
" ",
"ing"
],
"19": [
"．",
"／",
"ing",
"：\n",
"<|im_end|>",
"｜",
" ",
"［",
"：\"",
" "
],
"20": [
"ing",
"schlü",
"：\"",
"／",
"�",
"tória",
"’à",
"：\n",
"：“",
"ckså"
],
"21": [
"<|im_end|>",
"ing",
"�",
"’à",
"：\n",
"⁺",
"schlü",
"tória",
"’é",
";\r\r\n"
],
"22": [
"<|im_end|>",
"ing",
"：\"",
"schlü",
";\r\r\n",
"⁺",
"、“",
"�",
"tória",
"：\n"
],
"23": [
"�",
"<|im_end|>",
"ing",
"、“",
"：\"",
"’à",
");\r\r\n",
";\r\r\n",
"：“",
"tória"
],
"24": [
"�",
"<|im_end|>",
"ing",
"’à",
";\r\r\n",
");\r\r\n",
"tória",
"、“",
"schlü",
"：\""
],
"25": [
"<|im_end|>",
"。\n\n",
"ing",
"：\"",
"’à",
"！\n\n",
"、“",
");\r\r\n",
";\r\r\n",
"。\n\n\n\n"
],
"26": [
"<|im_end|>",
"tória",
" Erotische",
"schlü",
"ing",
");\r\r\n",
" lesbische",
"ckså",
";\r\r\n",
" erotisk"
],
"27": [
"<|im_end|>",
" erotisk",
" Erotische",
" tjejer",
" lesbische",
"tória",
"。",
"<|endoftext|>",
" lesbische",
"精准",
"。",
"<|endoftext|>",
" lesbische",
"。",
" lesbische",
"\"",
"tória",
"_corners",
");\r\r\n",
"。",
"<|endoftext|>",
"\"/>&#",
"');?>",
"<|endoftext|>",
" lesbische",
"(edges",
"tória",
"[offset",
"⁺",
"\"description",
"(messages",
"_matched"
],
"33": [
"<|im_end|>",
" lesbische",
"\"",
"]&#",
"(messages",
" предложения",
"[offset",
"\"description"
],
"34": [
"<|im_end|>",
" предложения",
" lesbische",
"摞",
"(messages",
"瘾",
"enen",
"\"description",
"收紧",
"ië"
],
"35": [
"<|im_end|>",
"摞",
" предложения",
" \n",
"瘾",
"<|endoftext|>",
"wargs",
"tória",
"$response",
"合伙人"
],
"36": [
"<|im_end|>",
"<|endoftext|>",
"摞",
"tória",
"mysł",
" предложения",
"Ẻ",
"收紧",
"шиб",
"$response"
],
"37": [
"<|im_end|>",
"\"text",
"шиб",
"ḃ",
"摞",
" أنح",
"$request",
"$response",
"局限",
"ayout"
],
"38": [
"\n",
"inders",
" предложения",
"lander",
"llib",
"铢",
"挤",
"句子",
"indered",
"(cards"
]
}
}
],
"trajectories": [],
"emergence": {
"position": 91,
"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": [
45485,
84834,
102804,
77111,
116269,
122959,
122128,
115221,
106114,
134346,
130471,
101829,
36641,
93589,
118683,
93678,
25323,
29097,
3593,
117220,
151670,
128412,
142402,
150159,
149443,
144434,
148475,
149323,
139870,
138026,
127742,
142244,
39337,
43724,
15489,
2085,
919,
30,
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."
}