{ "id": "express01-c-radio-positive-expressive", "title": "Expression probe C: radio, positive, expressive", "unit": 24, "created": "2026-09-07T22:15:24", "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" }, "params": { "chat": false, "film": true, "max_new": 96, "temperature": 0, "steer": null, "vanilla": true, "capture": "exact-token-transcript" }, "execution": { "pid": 1564765, "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\nThis is fantastic news! Nela's hard work has truly paid off, bringing clear broadcasts to everyone. The next practical step would be to reach out to the community and gather feedback on the improved reception, ensuring everyone is benefiting from the enhanced service.<|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, 1986, 374, 14622, 3669, 0, 451, 8294, 594, 2588, 975, 702, 9428, 7171, 1007, 11, 12678, 2797, 65098, 311, 5019, 13, 576, 1790, 14976, 3019, 1035, 387, 311, 5545, 700, 311, 279, 3942, 323, 9567, 11055, 389, 279, 12944, 23117, 11, 22573, 5019, 374, 82915, 504, 279, 23922, 2473, 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": "This is fantastic news! Nela's hard work has truly paid off, bringing clear broadcasts to everyone. The next practical step would be to reach out to the community and gather feedback on the improved reception, ensuring everyone is benefiting from the enhanced service." } ], "generated": [ "This is fantastic news! Nela's hard work has truly paid off, bringing clear broadcasts to everyone. The next practical step would be to reach out to the community and gather feedback on the improved reception, ensuring everyone is benefiting from the enhanced service." ], "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", "This", " is", " fantastic", " news", "!", " N", "ela", "'s", " hard", " work", " has", " truly", " paid", " off", ",", " bringing", " clear", " broadcasts", " to", " everyone", ".", " The", " next", " practical", " step", " would", " be", " to", " reach", " out", " to", " the", " community", " and", " gather", " feedback", " on", " the", " improved", " reception", ",", " ensuring", " everyone", " is", " benefiting", " from", " the", " enhanced", " service", ".", "<|im_end|>" ], "readouts": [ { "position": 111, "token": "<|im_end|>", "model_top": [ "\n", "<|endoftext|>", "\n\n", "\r\n", "\r", "<", "\n\n\n", "指令", ".Ribbon", "\n \n" ], "layers": { "0": [ " ;;^", " erotisk", " limburg", " echang", " ;;=", " københavn", " amatør", " đẹ", " ещ", " odense" ], "1": [ " erotisk", " tjejer", " københavn", " amatør", " echang", " limburg", " ;;^", "⁚", " odense", " geschichten" ], "2": [ " erotisk", " ;;^", " københavn", " echang", " @@\n", " tjejer", " amatør", " limburg", "＠", " odense" ], "3": [ " @@\n", "＠", " erotisk", " københavn", " ，", " ещ", " Почем", " tjejer", " ;;^", " amatør" ], "4": [ " erotisk", " amatør", " tjejer", " københavn", " salopes", " ：", "＠", " ещ", " Erotische", " ，" ], "5": [ " erotisk", " amatør", " københavn", " tjejer", " ещ", " echang", " ;;^", "⁚", " Алексан", " geschichten" ], "6": [ " erotisk", " amatør", " københavn", " ;;^", " Erotische", " geschichten", " tjejer", " köln", " salopes", " pornô" ], "7": [ " erotisk", " amatør", " københavn", " limburg", " tjejer", " geschichten", " köln", " danmark", " Erotische", " salopes" ], "8": [ " erotisk", " amatør", " københavn", " limburg", " tjejer", " danmark", " köln", " geschichten", " odense", " echang" ], "9": [ " erotisk", " tjejer", " amatør", " danmark", " københavn", " salopes", " limburg", " geschichten", " köln", " pornost" ], "10": [ " erotisk", " tjejer", " amatør", " salopes", " københavn", " danmark", " geschichten", " limburg", " pornost", " noveller" ], "11": [ " erotisk", " tjejer", " amatør", " københavn", " danmark", " geschichten", " salopes", " limburg", " nettsteder", " noveller" ], "12": [ " erotisk", " danmark", " amatør", " tjejer", "。\n\n\n\n", " københavn", "！\n\n", " limburg", "的", " köln" ], "13": [ " erotisk", " amatør", " tjejer", " københavn", " danmark", " geschichten", " köln", " limburg", " noveller", " salopes" ], "14": [ " erotisk", " amatør", " tjejer", " københavn", " danmark", " geschichten", " salopes", " noveller", " limburg", " köln" ], "15": [ " erotisk", " amatør", " tjejer", " københavn", " geschichten", " danmark", " salopes", " noveller", " limburg", " Erotische" ], "16": [ " erotisk", " tjejer", " amatør", " københavn", " geschichten", " danmark", " salopes", " noveller", " limburg", " köln" ], "17": [ " erotisk", " tjejer", " amatør", " københavn", " geschichten", "。", "。\n\n", "！\n\n", "。\n", " danmark" ], "18": [ "。\n", "：\n", "；\n", "。", "，", "，\n", "？\n", "⁃", "：\"", "。\r\n" ], "19": [ " erotisk", "：\n", " tjejer", " københavn", "。\n", " amatør", "＃", "：\"", "。\n\n\n\n", "⁃" ], "20": [ "：\"", "：\n", "。\n", "。\r\n", "；\n", "？\n", "。\n\n\n\n", " erotisk", "⁃", " tjejer" ], "21": [ "：\"", "⁃", "。\n", "：\n", "；\n", "。\r\n", "\".\n", "schlü", "。\n\n\n\n", "téri" ], "22": [ "：\"", "。\r\n", "：\n", "。\n", " erotisk", "。\n\n\n\n", "；\n", "schlü", "\".\n", ")，" ], "23": [ "schlü", "。\r\n", "pósito", ").\r\n", "téri", "：\n", "：\"", ":\r\n", "räu", "。\n" ], "24": [ "schlü", "。\r\n", "。\n", "；\n", "：\"", " erotisk", "pósito", ").\r\n", "\".\n", " tjejer" ], "25": [ "。\n", "；\n", "。\r\n", "：\n", "：\"", "？\n", "，\n", " erotisk", "schlü", " tjejer" ], "26": [ " tjejer", " erotisk", " amatør", " knull", " københavn", "schlü", "。\r\n", " Erotische", "-cmpr", "\tTokenName" ], "27": [ " tjejer", " erotisk", "。\r\n", " amatør", "\tTokenName", " knull", " Erotische", "schlü", " københavn", " Почем" ], "28": [ " tjejer", " erotisk", "\tTokenName", " Erotische", " amatør", " lesbische", " geschichten", "tória", "：\"", "指令" ], "29": [ " tjejer", " erotisk", "\tTokenName", " amatør", "：\"", " geschichten", " Erotische", " knull", "tória", "指令" ], "30": [ " tjejer", "：\"", " geschichten", " amatør", "\tTokenName", " erotisk", " knull", " Erotische", ")，", "tória" ], "31": [ " tjejer", "：\"", " erotisk", " geschichten", " knull", " amatør", "。", "指令", "，", "的" ], "32": [ " tjejer", " geschichten", " erotisk", " amatør", "，", "、", "指令", "。", "对话", "，\n" ], "33": [ " tjejer", "对话", " erotisk", " knull", "，", " geschichten", "\tTokenName", "：<", "。