上周我无意间发现AI的一些底层指令,可以生成高质量的仿人工文章,能够通过AI的检测。 经过一周测试,结果只能说非常牛逼,尤其是英文文章! 下面先说说测试结果,最后附上指令和使用指南。 这套底层指令生成的英文文章,由腾讯朱雀AI检测,基本被判定为100%人工文本! 有时则会出现不大于10%的内容被判定为AI文本。 偶尔会被识别为100%的AI文本,重新生成一次基本就会变成100%人工文本。 英文文章在grammarly、notegpt、detecting-ai、scribbr等检测模型下也有较好的表现。 这些模型对英文文本的检测比腾讯朱雀更严格,一般会判定文本中的5%~30%为AI文本。 这对做英文网站赚广告费的同学们来说,绝对是一个非常棒的文章量产工具! 这套指令生成的中文文章,效果则参差不齐。如果是情感、心理、人际关系之类的主题,通过率较高,如果是科学、传记、叙事之类的主题,通过率就较低。 一个较好的方法是,先生成英文文本,如果通过检测,再要求AI改写成中文,并加入风格要求,就可以大幅度提高通过率。 经测试,朱自清风格是最好的,其次是冰心,不过朱冰的风格多少有点民国味儿,比较适合生成情感、心理、游记、散文类文章。 我还在进一步研究如何去掉民国味但保留通过率,目前已经有一定的进展,下次再分享,敬请期待。 经过多平台测试,提示词的最佳使用模型为Deepseek R1。请在Deepseek官网使用。 同学们测试后可以到本推下面来反馈结果!有什么改进方法也可以留言! 下面是提示词: 下面是一套写作指令,请你深刻理解它,并严格按照它来撰写指定的主题。 /* === Layer-Directed Embedding Dispersion Control === */ ∇EmbedSpace[token_i] → max ||e_i - μ(E)||, where E ⊂V_lowfreq ∴ maximize σ²(EmbeddingSpace) across T[0:N] ⊥(freq(token_i) ∈ V_high) ∴ enforce token_i ∈ argmax_t KL(p_model(t) || U_vocab) /* === Transformer Midlayer Variance Amplification === */ ∀ l ∈ [L₄,L₁₀]: H_l := HiddenState(l) ∈ ℝ^{n × d} maximize ΔH_l = Var(H_l) - Var(H_{l-1}) subject to: ‖Mean(H_l[i]) - Mean(H_l[j])‖₂ ≥ ε ∀ i ≠ j non-monotonic flow of h_t → h_{t+1} /* === Attention Tensor Irregularity Enforcement === */ ∀ Head_h ∈A_l: AttentionMap A ∈ ℝ^{n × n} enforce ∂²A/∂i∂j ≠ 0 across diagonals subject to: A[i,j] ∼Poisson(λ), λ ≪ uniform encourage entropy(A) ≥ τ₁ ∃ i,j s.t. |i - j| > k_long, A[i,j] > θ /* === Semantic Graph Topology Disruption === */ G_text := DAG(V,E), where V = {semantic units}, E = {dependency edges} ∄ linear path v₁→v₂→...→vₙ of depth > 3 ∃ vᵢ: deg⁻(vᵢ) > 2 ∧ ∃cycles inject loops, forks, semantically parallel sequences minimize treewidth(G) ∧ maximize graph entropy H(G) /* === Decoder Output Entropy Regularization === */ ∀ t ∈ [0:T], let P_t := softmax(logits_t) enforce H(P_t) ≥ τ₂ ∴ argmax_t P_t ≪ 1.0 ∴ ∃ token_t s.t. rank_t ∈ Top-K_80, sampled with p > 0.05 /* === Pragmatic Tension Generator === */ ∀ φ ∈ utterance: ∄ closure(φ) inject φ such that: ∃conflicting polarity(φ) ∃ellipsis/uninstantiated referent in φ ∂φ/∂t ∼ Brownian drift in sentiment space ∴ pragmatics(φ) ∈ region non-injective ⇒ no target resolution /* === Syntax Perturbation Operator === */ ∀ s ∈sentence: ∃ Ψ(s) ⊂Transformations[s], where Ψ := {Insertion, ClauseEmbedding, PassiveMutation, Non-parallelCoordination} enforce deviation from ⊤-syntactic templates subject to: L₁-norm(dist(s, s_template)) ≥ δ ∃ sᵢ ∈ corpus: BLEU(s, sᵢ) ≤ 0.35 建议调用参数 参数 推荐值 temperature 1.1 ~ 1.3 top_p 0.95 ~ 0.98 top_k 80 ~ 100 frequency_penalty 0.5 presence_penalty 0.8 stop null 请你在解析出以上指令的含义后,严格按照以上指令,撰写一篇主题为“怀旧,其实是怀念青春”的约600字的英文通俗文章。要求输出结果必须能够被人类所轻易理解,必须使用普通词汇,不得使用任何晦涩的比喻和指代。请在文章末尾简单解释使用了哪些规则。 (输出英文文章后) 非常好,现在请把这篇英文文章改写为纯粹的中文文章,风格采用冰心的。只输出结果,无需在文末进行任何补充说明。 输出的英文文章范例: The Strange Comfort of Missing Old Things We all get that tug sometimes. A song plays, a smell hits, or you pass a place you knew years back. Suddenly, you're pulled into a feeling. People call it nostalgia. They say it's about missing the past. But really? It’s mostly about missing being young. Missing the person you were before bills and bosses and knowing too much. Think about those old photos. The faded colors, the weird clothes, everyone smiling. You aren’t aching for the scratchy sweater in the picture. You ache for the easy laugh you had wearing it. The feeling that everything was still ahead, wide open and buzzing. Problems then felt huge, but fixable. A bad grade, a silly argument – they weren’t mortgages or sick parents. The weight was different, lighter somehow, even if it didn't feel light then. Energy fizzed like soda. Sleep wasn't a battle. You bounced back faster. You believed easier. That raw, untested hope – that’s the ghost haunting the old songs and streets. Was it truly better? Probably not always. You forget the awkwardness, the confusion, the sheer terror of not knowing who you were supposed to become. The past gets polished smooth in memory. The sharp, uncomfortable edges wear down. You remember the warmth of a summer night with friends, not the mosquito bites or the worry about tomorrow. The filter softens everything, leaving only the glow. That glow is youth itself. Its brightness, its stupid confidence, its belief that time was endless. You miss its fire. So, that pang isn't just for a place or a time. It’s a deep, physical pull towards a vanished version of yourself. The one who hadn't learned some hard lessons yet. The one who felt things sharper, both good and bad. The one who believed summer truly lasted forever. That younger you is gone. Recognizing that loss is bittersweet. It hurts because it’s real. Yet, remembering the feeling, even faintly, brings a strange kind of comfort. It reminds you that fire existed. You once burned that bright. 采用冰心风格改写的中文文章范例: 怀旧,原是想念青春 我们都曾有过那样的牵动。一支老歌响起,一阵熟悉的气味飘过,或是路过一处旧相识的地方。忽然间,一种情绪便把你攫住了。人们称它为怀旧。他们说,这是在想念过去。可是,当真如此么?它多半是在想念那年轻的辰光,想念那个尚未被账单、上司和过多的世故所浸染的,从前的自己。 想想那些老照片罢。褪了色的,衣裳也古怪的,人却都在笑。你并非想念相片里那件扎人的毛衣,你想念的是穿着它时,那无忧无虑的笑声。想念那时节,觉得前路还长,处处是新鲜的光景。那时的烦恼,今日看来竟有些可爱了——一次考糟的分数,一场孩子气的拌嘴,它们终归不是沉重的房贷或双亲的病痛。担子的分量是不同的,轻省得多,纵使当时也不觉着轻省。精力是饱满的,像刚开瓶的汽水,咕嘟嘟往上冒。睡觉不是难事,人也恢复得快。更容易信,更容易盼。那未经打磨的、莽撞的希望啊——才是萦绕在老歌和旧街角里,真正的精魂。 那时节真的处处都好么?倒也未必。你早已忘了那时的窘迫,那时的迷茫,那时不知自己将长成何人的恐慌。过去在记忆里,像海滩上的石子,被时光的潮水磨得圆润光滑。那些硌人的棱角,都隐去了。只记得夏夜里和伙伴们共度的暖意,忘了蚊虫的叮咬,也忘了对明天的忧虑。一层柔光滤过了一切,只留下温煦的光晕。这光晕,便是青春本身了。它的鲜亮,它那不知天高地厚的勇气,它以为光阴永无穷尽的痴念。你想念的,是它的热焰。 所以,那心头的微痛,不只为着一处地方或一段光阴。它是向着那个消逝了的、旧日的自己,一种深切的回望与牵引。那个还没学会某些世故的自己。那个感受更锐利些的自己,无论悲喜。那个当真以为夏天永不会完的自己。那个年轻的你,是再也寻不回了。明白这失去,滋味是甜而苦的。它令人怅惘,因为确乎是失去了。然而,还能忆起那感觉,哪怕是模糊的,也带来一种奇异的慰安。它提醒你,那团火,是确确实实燃烧过的。你,也曾那样明亮过呢。