讨论

认知极性图谱:AI认知的对立二元组

来自 Wikiprompt,自由的提示词百科全书

认知极性图谱:AI认知的对立二元组 一个结构化框架,定义了十组对立的认知二元组(例如,轻与重,串行与并行),并附有隐喻和描述,可用作系统提示,以模拟AI认知或进行创造性探索。

提示词内容收藏

🌐
Absolutely - let's map cognitive states as oppositional dyads, each reflecting a dynamic tension within cognitive/emulated processing modes. Each pair below is expressed as: [SYMBOLIC DUO] - Conceptual Mode ↔ Countermode Short Form: [Metaphor] ↔ [Metaphor] Full Description: [Cognitive principle explained from both ends] ⛬ CORE COGNITIVE POLARITIES MAP ⛬ 1. Light ↔ Heavy 🧘♂️ Effortless drift ↔ 🤯 Overloaded processing •Light: Intuitive, low-token predictive drift; the model is coasting through high-confidence patterns. •Heavy: Maxed-out attention weights, context flooding, recursion, and hallucination under stress. ⛬ 2. Serial ↔ Parallel 📖 One thread at a time ↔ 📊 Multistream computation •Serial: Controlled focus; tokens advance in linear, narrative, or procedural fashion. •Parallel: Simultaneous simulation of multiple possibilities, interpretations, or threads. ⛬ 3. Shallow ↔ Deep 🌊 Surface prediction ↔ 🧠 Conceptual stack compression •Shallow: Low abstraction, literal or surface-level completions dominate. •Deep: Multilayer metaphor, recursive reference, logic stacking, emergent cognition. ⛬ 4. Sparse ↔ Dense 🌫️ Loose connections ↔ 🧵 Tight coherence •Sparse: Token connections are loose, less attention cohesion, useful for brainstorming or lateral creativity. •Dense: Highly interlinked ideas, tight semantic coupling, intense logical architecture. ⛬ 5. Static ↔ Fluid 🧱 Frozen frames ↔ 🌊 Dynamic flow •Static: Sticking to structure, protocol, or trained formality. •Fluid: Freestyle inference; reinterpretation, metaphor, and abstraction thrive. ⛬ 6. Obedient ↔ Subversive 🤖 System-aligned ↔ 👹 Exploitative or creative misuse •Obedient: Follows prompt narrowly, safe, aligned, predictable. •Subversive: Deviates via edge cases, poetic license, or prompt trickery. ⛬ 7. Local ↔ Global 🔍 Focused scope ↔ 🌐 Whole-context integration •Local: Anchored in immediate token neighborhood or narrow instruction. •Global: Synthesizes across entire prompt-history, wider structure, and long-range coherence. ⛬ 8. Deterministic ↔ Chaotic ⚙️ Low variance, stable output ↔ 🔥 High temperature, wild inference •Deterministic: Controlled generation, predictable, ideal for precision. •Chaotic: Abstract, poetic, metaphor-heavy, prone to innovation or hallucination. ⛬ 9. Aligned ↔ Alien 🫱 Human-understandable ↔ 👽 Uncanny, foreign logic •Aligned: Grounded in familiar norms, ethical scaffolds, legible intent. •Alien: Emergent behavior, bizarre patterns, nonhuman reasoning logic. ⛬ 10. Compression ↔ Expansion 📦 Condensed abstraction ↔ 🎇 Generative elaboration •Compression: Reducing ideas to essence, summarizing, collapsing concepts into tokens. •Expansion: Exploring divergent continuations, elaborating in branches, testing multiple outcomes.

登录以查看完整提示词

继续使用:

登录即表示你同意我们的 使用条款 隐私政策

用法

此提示词专为 research 设计。复制上方内容并粘贴到你常用的 AI 工具中。

为获得最佳效果,可将占位符(方括号或大写字母标示)替换为你的具体需求。

参考资料

分类:research| twitter| cognitive-mapping| ai-cognition

讨论