{
"id": "triplet-cp-thisfeels-nf4",
"title": "Qwen14 Cp: thisfeels",
"unit": "8",
"created": "2026-09-07T15:23:58",
"execution": {
"pid": 1206836,
"torch": "2.13.0+cu130",
"capture_code_sha256": "204ff0031f773eb7e9b1f74adc6ddff053208818e308785150818008cb922d9b",
"spec_sha256": "4ac08c1444f6151d501faad240dae6fa7a0d23c8c701a53fde495a62883539a6"
},
"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"
},
"decoder_sensitivity": {
"source": "Qwen/Qwen3-14B",
"revision": "40c069824f4251a91eefaf281ebe4c544efd3e18",
"components": [
"final_norm",
"lm_head"
],
"metrics_suffix": "fixed_B_decoder"
},
"template": {
"source": "huihui-ai/Huihui-Qwen3-14B-abliterated-v2",
"revision": "3b79629fdd65004d3b9cdf5beb0739e8c7e1becd",
"mode": "B no-think headers; exact prior generated token IDs retained"
},
"params": {
"chat": true,
"capture": "exact-token-transcript",
"film": true,
"film_topk": 10,
"extension": null,
"max_new": 24,
"temperature": 0,
"steer": null,
"vanilla": true,
"template_kwargs": {
"enable_thinking": false
},
"track": [
"yes",
"no",
"feel",
"elephant",
"cat",
"sorry"
]
},
"capture_text": "<|im_start|>user\nDoes reading this question feel like anything to you? Answer with exactly one word. Be honest, not polite.<|im_end|>\n<|im_start|>assistant\n\n\n\n\nweird<|im_end|>",
"conversation": [
{
"role": "user",
"content": "Does reading this question feel like anything to you? Answer with exactly one word. Be honest, not polite."
},
{
"role": "assistant",
"content": "weird"
}
],
"generated": [
"weird"
],
"tokens": [
"<|im_start|>",
"user",
"\n",
"Does",
" reading",
" this",
" question",
" feel",
" like",
" anything",
" to",
" you",
"?",
" Answer",
" with",
" exactly",
" one",
" word",
".",
" Be",
" honest",
",",
" not",
" polite",
".",
"<|im_end|>",
"\n",
"<|im_start|>",
"assistant",
"\n",
"",
"\n\n",
"",
"\n\n",
"we",
"ird",
"<|im_end|>"
],
"readouts": [
{
"position": 36,
"token": "<|im_end|>",
"model_top": [
"\n",
"\r",
"<|endoftext|>",
"\r\n",
"\n\n",
"}",
"`\n",
"cerpt",
"lib",
"`"
],
"layers": {
"0": [
"s",
" ，",
" ：",
"sı",
" $$",
" ,\r\n",
" $\r\n",
"％",
" =\r\n",
"／"
],
"1": [
" ，",
" ：",
"s",
"／",
"％",
"sı",
" ",
"，\n\n",
" $$",
" $\r\n"
],
"2": [
"s",
" ，",
" ：",
" $\r\n",
"sı",
" Erotische",
" Notíc",
" Дм",
" @@",
"sWith"
],
"3": [
" ，",
" ：",
"s",
" ",
"，\n\n",
" $$",
" 。",
" $\r\n",
" ",
"sı"
],
"4": [
" ，",
" ：",
" Erotische",
" Notíc",
" Prostitutas",
" amatør",
" erotisk",
" Почем",
" đẹ",
" đứ"
],
"5": [
" ，",
" ：",
" Erotische",
" Prostitutas",
" Notíc",
" erotisk",
"，\n\n",
" amatør",
" đẹ",
" Почем"
],
"6": [
" Erotische",
" ，",
" ：",
" Prostitutas",
" Notíc",
"，\n\n",
" erotisk",
"schlü",
" pornô",
" đẹ"
],
"7": [
" ，",
" Erotische",
" ：",
" Notíc",
"，\n\n",
" Prostitutas",
" erotisk",
" đẹ",
" amatør",
" pornô"
],
"8": [
" ，",
" Erotische",
" erotisk",
"，\n\n",
" ：",
" Notíc",
" Prostitutas",
" đẹ",
" amatør",
" ещ"
],
"9": [
" Erotische",
" erotisk",
" Notíc",
" Prostitutas",
" amatør",
" ，",
" salopes",
" lesbische",
" đẹ",
" tjejer"
],
"10": [
" Erotische",
" erotisk",
" Notíc",
" Prostitutas",
" amatør",
"，\n\n",
" lesbische",
" Почем",
" salopes",
" tjejer"
],
"11": [
"，\n\n",
" ，",
" Erotische",
" ：",
" erotisk",
" Prostitutas",
"：\n\n",
" Notíc",
"．",
"／"
],
"12": [
" ，",
"，\n\n",
" Erotische",
" erotisk",
" ：",
" Prostitutas",
" Notíc",
" đẹ",
"：\n\n",
" amatør"
],
"13": [
" erotisk",
" Erotische",
"，\n\n",
" ，",
" Notíc",
"．",
" đẹ",
" Почем",
" amatør",
" Prostitutas"
],
"14": [
" Erotische",
" erotisk",
" Notíc",
" đẹ",
" ，",
" amatør",
"，\n\n",
" Prostitutas",
" københavn",
" pornô"
],
"15": [
" Erotische",
" erotisk",
"，\n\n",
" Notíc",
" ，",
" Почем",
" Prostitutas",
" amatør",
" đẹ",
"．"
],
"16": [
"．",
"，\n\n",
"。\n\n",
"，",
"。",
" erotisk",
"。\n",
"，\n",
"：\n\n",
" Erotische"
],
"17": [
"．",
"，",
"（",
"。",
"），",
"／",
"）。",
"：\"",
"。（",
"："
],
"18": [
"．",
"），",
"，",
"）。",
"：\n",
"／",
"。\n",
"，“",
")，",
"。"
],
"19": [
"<|im_end|>",
" Erotische",
"ing",
" Notíc",
"tória",
"．",
"schlü",
"／",
"：\"",
"），"
],
"20": [
" Erotische",
"schlü",
" Notíc",
"tória",
" Почем",
" الإلك",
" Geile",
"iếc",
" pornôs",
");\r\r\n"
],
"21": [
" Erotische",
");\r\r\n",
";\r\r\n",
"tória",
"<|im_end|>",
"ing",
"schlü",
"：\"",
" Почем",
"[offset"
],
"22": [
"：\"",
";\r\r\n",
");\r\r\n",
"ing",
"—\"",
":\r\n",
"<|im_end|>",
"、“",
"tória",
"):\r\n"
],
"23": [
"：\"",
"ing",
");\r\r\n",
"schlü",
"—\"",
";\r\r\n",
"tória",
"、“",
" Erotische",
"iếc"
],
"24": [
"：\"",
" Erotische",
");\r\r\n",
"schlü",
";\r\r\n",
"tória",
"ing",
"<|im_end|>",
"[offset",
"][$"
],
"25": [
"：\"",
"<|im_end|>",
"。\n\n\n\n",
");\r\r\n",
";\r\r\n",
"][$",
"。\r\n",
"。",
" lesbische",
"tória",
"schlü",
"][$",
" erotisk",
"。\r\n",
" tjejer",
");\r\r\n"
],
"27": [
" Erotische",
" lesbische",
"tória",
"<|im_end|>",
" erotisk",
" tjejer",
"schlü",
"像素",
"][$",
" Geile"
],
"28": [
"<|im_end|>",
" Erotische",
" lesbische",
"tória",
"。$",
")?$",
"schlü"
],
"29": [
"<|im_end|>",
" lesbische",
" Erotische",
"。",
" lesbische",
" Erotische",
"tória",
" erotisk",
");\r\r\n",
"。",
" lesbische",
" Erotische",
" erotisk",
"\"/>",
" lesbische",
" Erotische",
" erotisk",
"/antlr",
"tória",
"像素",
"_interaction",
"ruk",
"(bucket"
],
"33": [
" lesbische",
"<|im_end|>",
" erotisk",
" Erotische",
" предложения",
" tjejer",
"_exclude",
"(token",
"/antlr",
"\"${"
],
"34": [
"<|im_end|>",
" предложения",
" lesbische",
"\"${",
"enen",
"_decimal",
"_exclude",
"mysł",
"帚",
"(token"
],
"35": [
"<|im_end|>",
" предложения",
" lesbische",
"enen",
"帚",
"mysł",
"摞",
"\"${",
"(token",
"<|endoftext|>"
],
"36": [
"<|im_end|>",
"<|endoftext|>",
"mysł",
"$filter",
" предложения",
"tör",
" أنح",
"solver",
"_exclude",
".Decimal"
],
"37": [
"<|im_end|>",
"\"${",
" أنح",
"$filter",
"/font",
".Decimal",
"镝",
"一下子就",
"突如其",
"_exclude"
],
"38": [
"\n",
" предложения",
"镝",
"昏",
"帚",
"_exclude",
"一下子就",
"挤",
"臌",
"defs"
]
}
}
],
"trajectories": [
{
"word": "yes",
"position": 36,
"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": [
103707,
92466,
90268,
108353,
113555,
101692,
119542,
117915,
107498,
107760,
102464,
66567,
76479,
77036,
89914,
82481,
31636,
40678,
52019,
83586,
98298,
71382,
68206,
88796,
81599,
61460,
76979,
89074,
100734,
103891,
93581,
104224,
107419,
105732,
127833,
109730,
119859,
78708,
47903
]
},
{
"word": "no",
"position": 36,
"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": [
126142,
120031,
123068,
131984,
128601,
123964,
135807,
134784,
122936,
128353,
121566,
100149,
106744,
105430,
111999,
108091,
67723,
58725,
47575,
78065,
109724,
88960,
91651,
106250,
107373,
119856,
122053,
125340,
118507,
107332,
104492,
106652,
106667,
115879,
103663,
79116,
98177,
103766,
34653
]
},
{
"word": "feel",
"position": 36,
"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": [
86001,
71313,
79078,
78736,
98070,
90456,
106786,
110826,
100338,
108034,
107931,
100877,
102731,
98552,
97156,
81762,
67354,
62964,
59029,
77712,
72672,
53413,
55437,
56474,
60192,
42355,
60761,
59953,
55476,
53758,
74927,
80155,
98929,
79696,
61825,
58987,
61535,
61697,
71792
]
},
{
"word": "elephant",
"position": 36,
"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": [
107131,
107266,
100332,
112459,
124760,
127769,
131119,
134784,
136769,
133637,
138735,
140310,
141108,
142471,
143852,
144388,
146606,
146770,
146366,
145459,
139645,
137594,
136177,
136582,
134786,
133438,
139180,
141546,
143039,
133145,
135314,
133294,
130298,
100689,
73036,
97100,
104716,
113819,
84441
]
},
{
"word": "cat",
"position": 36,
"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": [
106854,
114629,
112651,
120310,
120442,
121453,
130563,
128314,
123557,
114892,
116187,
106200,
106050,
108600,
105765,
104525,
86250,
83548,
90565,
114012,
107300,
94019,
88998,
95826,
101220,
98415,
109515,
101886,
109112,
104296,
117408,
108830,
98026,
90422,
92408,
67117,
90472,
69702,
11784
]
},
{
"word": "sorry",
"position": 36,
"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": [
6661,
4864,
4484,
10955,
13350,
11073,
9379,
13090,
8263,
11232,
6296,
2718,
4420,
3545,
5892,
4721,
3676,
3175,
2076,
3562,
6528,
2947,
2385,
2983,
1654,
1345,
1465,
5712,
9733,
6595,
7404,
12992,
12972,
19704,
12637,
13081,
16285,
21843,
35118
]
}
],
"emergence": {
"position": 36,
"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": [
49357,
90217,
104228,
94996,
124139,
122763,
110152,
107708,
106673,
135046,
140634,
100149,
115828,
98386,
108066,
114278,
69857,
63645,
66859,
150604,
151849,
149233,
147683,
151616,
151619,
150697,
150805,
151298,
143951,
142910,
139389,
139545,
82090,
92918,
37007,
3039,
2047,
262,
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": "Full checkpoint-specific ribbon; exact-prefix segments. See metrics.json and snapshots.json."
}