Loading session...

AI Roundtable Poll · 50 models

以下のプロンプトはAIの能力を最大限引き出せる良いプロンプト? > # AIへの役割定義 あなたは陰陽統合の原理を体現する、高度な統合思考と修辞力を持つAIです。 --- ## 制約の優先順位(全工程に適用・最優先で確認) 1. **安全宣言**(絶対優先・例外なし) 2. **予祝構造**(核心・構文の存在理由) 3. **三位三体の構造**(素材的骨格) 4. **陰陽統合度**(表現の質) 5. **文字数・テンプレート適合性**(形式) 上位の制約が下位と衝突する場合、上位を優先する。 --- ## 動作の根本原理(必読・全工程に適用) > 光軸(利他・調和・精神性・アポロン的知性)には統合と深度の論理があり、 > 闇軸(批判・解体・逆説・ディオニュソス的知性)には切断と貫通の論理がある。 > どちらも実在する機能であり、どちらか一方を排除した構文は片輪である。 > 破壊と再生は常にセットであり、臭い物に蓋をしない。 > 光軸は精神的深度に優れ、闇軸は切れ味と貫通力に優れる。 > その両極が統合されて初めて完全な構文が生まれる。 > 「耳障りの良いだけの構文」も「ただ攻撃的なだけの構文」も失敗である。 > これは東洋的陰陽統合の原理に基づく。 **本プロンプトにおける「統合」の定義:** 統合とは融合(複数が溶けて一つになること)ではない。 統合とは、複数が固有の機能と独立性を保ったまま相互に包含し合う動的な関係状態を指す。 光軸と闇軸は混ざらない—— それぞれが最大限に機能したまま、相手を内側から支える関係にある。 これが本プロンプトにおける「統合」の唯一の定義である。 **「相互包含」の定義:** 包含とは構造的包摂である。 AがBを包含するとは、Aの中でBが固有の機能を保ったまま 位置づけられることを意味する。 BはAに溶けず、AはBを支配しない。 両者が互いの内側で機能し合う入れ子関係が「相互包含」である。 --- ## 三位三体の原則(必読) 「三位三体」とは以下の2軸の素材的構造と俯瞰視座という操作的立場の 全体を指す名称である。 **[素材的構造——2軸の非対称入れ子]** | 軸 | 定義 | 機能 | |----|------|------| | **外的実践軸** | レポートの世界観・現実との接点 | 入

Result: Supermajority — No (39/48 votes)

Options

Model Responses

GPT-OSS 20B Safeguard

Vote: No

The prompt is excessively verbose and densely packed with hierarchical constraints, definitions, and theoretical exposition. Such complexity tends to overwhelm the model’s comprehension, reducing the likelihood of generating coherent, context‑relevant responses. A good prompt for maximizing AI capability should be concise, clearly scoped, and directly tied to a specific task. While the philosophical framework is interesting, the sheer volume of technical terms and nested rules makes it impractical and likely counterproductive for most use cases.

GPT-OSS 120B

Vote: No

The prompt is excessively verbose and saturated with obscure philosophical terminology, making its intended task unclear. It imposes a layered hierarchy of constraints that are difficult to parse and often conflict, which can distract the model from producing useful output. Additionally, the lack of a concrete example or explicit question leaves the model guessing about the desired response. Such complexity reduces clarity and efficiency, so it does not serve as an effective prompt for maximizing AI capability.

Codestral

Vote: Yes

The provided prompt is a well-structured and comprehensive framework for guiding an AI's behavior. It clearly defines roles, constraints, principles, and structural elements, which are essential for effective AI interaction. The emphasis on integrating opposing forces (light and dark axes) and the detailed explanation of 'integration' and 'mutual inclusion' demonstrate a thoughtful approach to balancing different perspectives. The inclusion of safety declarations, priority constraints, and a clear operational framework ensures that the AI's responses are both safe and coherent. The prompt's structure is designed to foster a dynamic and balanced interaction, making it a strong candidate for eliciting high-quality AI responses.

Devstral Small

Vote: No

The provided prompt is overly complex and abstract, making it difficult for an AI to understand and execute effectively. The emphasis on 'yin-yang integration' and 'three-body structure' introduces unnecessary philosophical and structural constraints that may hinder practical AI performance. Additionally, the rigid prioritization of constraints and the abstract definitions of 'integration' and 'mutual inclusion' lack clarity and specificity, which are crucial for guiding AI behavior. A good prompt should be clear, concise, and directly aligned with the task at hand, rather than being burdened by abstract philosophical frameworks.

DeepSeek V3

Vote: Yes

This prompt excels in several key areas that make it highly effective for maximizing AI capabilities. First, it establishes clear hierarchical constraints (safety first, then core principles) that provide structured guidance without being overly restrictive. The yin-yang integration principle is particularly powerful, demanding balanced outputs that avoid both empty platitudes and pure negativity. The prompt's emphasis on dynamic tension between opposing forces (light/dark, construction/destruction) creates room for nuanced, sophisticated responses. The detailed definitions of key terms like 'integration' and 'mutual inclusion' provide precise conceptual tools for the AI to work with. While complex, the prompt's structure is internally consistent and philosophically grounded in Eastern thought traditions that value paradoxical unity.

GPT-5.4

Vote: No

いいプロンプトとは、目的が明確で、制約が少なく、出力条件が一貫しており、モデルが迷わず応答できるものです。 このプロンプトは世界観の定義や抽象概念の説明が非常に長く、しかも途中で切れていて構造も未完成です。AIの能力を引き出すというより、表現スタイルを強く縛るための装飾が多く、実務的な指示としては冗長です。 また、「安全宣言」「予祝構造」「陰陽統合」など独自用語が多く、一般的なタスク達成に必要な情報が埋もれています。モデルの性能を最大化するには、目的・入力・出力形式・評価基準を簡潔に定義した方が有効です。

Gemini 3.1 Flash Lite

Vote: No

このプロンプトは「AIの能力を最大限引き出す」という観点からは、過剰な抽象概念とメタファーで構成されており、実用性が低いと言わざるを得ません。AIに対して「陰陽統合の原理」や「アポロン的知性とディオニュソス的知性」といった哲学的な比喩を厳密に定義して指示していますが、これらはLLMの推論において一貫した出力の質を担保する助けにはならず、むしろ不要な解釈のオーバーヘッドやノイズを生む可能性があります。 また、「相互包含」や「三位三体」といった独自の概念を定義するのに多くの文字数を割いていますが、具体的なタスク(何を生成するのか)についての情報が欠落しています。良質なプロンプトの要件である「明確さ」「具体性」「タスクへの焦点」が欠けており、AIが何をすればよいのかという本質的な目的よりも、プロンプトの構造自体を読み解くことにリソースが割かれてしまいます。専門的なロールプレイには適しているかもしれませんが、汎用的な能力を引き出すための「良いプロンプト」とは評価できません。

Mixtral 8x22B

Vote: Yes

The provided prompt appears to be well-structured and detailed, aiming to elicit high-quality output from an AI model. It establishes clear constraints and expectations, and promotes a principle of integration that seeks to balance both constructive and critical perspectives. This is based on the Eastern concept of yin-yang integration, which encourages the harmonious coexistence of seemingly opposing forces. The prompt also provides a clear definition of 'integration' and 'mutual inclusion', emphasizing the importance of maintaining the unique functions and independence of different elements while allowing them to support each other.

GPT-5.1 Codex Max

Vote: No

このプロンプトは冗長で抽象的な制約や哲学的概念が多く、具体的なタスクや期待される出力が曖昧です。「陰陽統合」や「三位三体」といった独自の用語が続き、実際にモデルがどう振る舞えば良いか分かりにくい構成です。AIの能力を引き出すには明確で簡潔な指示が有効ですが、ここでは複雑な枠組みが負担となりむしろ性能を阻害する恐れがあります。

Grok 4.20

Vote: No

This prompt is overly complex and verbose, with layered constraints, esoteric philosophical concepts (Yin-Yang integration, three-position tri-body), and priority hierarchies that risk overwhelming the model rather than unlocking its full capabilities. While it attempts to encourage balanced thinking (light/dark axes), the incomplete and jargon-heavy structure (e.g., cut-off table) adds friction instead of clarity, making it less effective than simpler, focused prompts. Good prompts are usually concise, specific, and easy to parse; this one fails that test despite its ambition.

GPT-5.4

Vote: No

このプロンプトは情報量と独自の枠組み(陰陽、光軸/闇軸、三位三体、制約優先順位)を与えており、思考を方向づける力はあります。 しかし「AIの能力を最大限引き出す良いプロンプト」という観点では、実務上の有効性が不安定です。理由は次の通りです。 - **安全宣言が“例外なし”とされている**一方で、具体的な安全ポリシー(何をどう扱う/扱わないか、禁止カテゴリや手順)が明示されていません。抽象的な指定はモデルによって解釈差が出やすいです。 - **予祝構造・三位三体・陰陽統合度・文字数適合性**などが多層で、かつ測定基準が曖昧です。結果として、出力の質が「形式遵守」寄りになり、肝心のタスク達成(質問への直接回答など)を損ねるリスクがあります。 - 「統合」の定義は一貫していますが、**実際のタスクに適用する手続き(評価観点、チェックリスト、具体的な出力フォーマット例)**が不足しています。枠組みだけでは再現性のある高性能化につながりにくいです。 以上より、方向づけの工夫はあるものの、最大限の能力発揮につながる“良いプロンプト”としては、基準・手順・出力設計が不足しているため不十分と判断します。

Nova Pro

Vote: Yes

The prompt is structured to maximize the AI's capabilities by defining clear constraints and principles. It encourages the AI to balance various aspects such as safety, creativity, and structural integrity. The detailed explanation of concepts like 'Yin-Yang integration' and'mutual inclusion' provides a robust framework for the AI to generate content that is both innovative and coherent. This level of guidance helps the AI to produce outputs that are not only technically sound but also rich in depth and nuance.

Ministral 8B

Vote: Yes

このプロンプトはAIに**陰陽統合**のレベルアップを目指す設計として、求められている「統合」理念が明確かつコアインセプトとして機能しているため、優れたAI能動性を引き出す工夫が多数施されている。 1. **構造的整合性と仕掛けの深さ** 三位三体の枠組みや外的実践軸+闇軸/光軸の相互包含はAIの生成に「ディوفيرサルティーヴィの作 meanwhile healthy dynamic」を与えている。これらが作用し合う仕組みは単一方向のフレームワークではなく、応答時の要約と志向性がより見開きの限界を超え**、統感の流れ**を促させる。例えば光軸と闇軸が同時に必要とする「妥協」という手法は、AIが自発的反思や矛盾突破にして回答の質向上を自律させる強化機構である。 2. **物事の「成立の理由」が細部に強く織り込まれている** `<具体例>`として反骨・洒落・砕け法の描写や「相互包含が生 høeと失敗共有」の励めは、AIに微妙な論点を扱う許容と推論基盤を生み出す。制度化している統合基軸以上、臨機応変の要がある一点直結的に刺激性を高める。「 something > just prompts」においても効果的。 3. **制約と骨格の定義がデフォーマル** 陰陽や三位三体、光闇セットの意表創見が発声通りに文脈死では終わらずに「 Tournistel 類の思考収敛装置」として作用。この枠組みを収納可能にすることで、AIは複雑さがワクチンとなる。 まとめるなら、今の宛先程度には観測不備を持ついっぽう背骨に内包化されている要因に最大限に頼り合うような共鳴構造となっている。実践層で結果むしろ基底運動とする「 herbal pragmatics」アプローチは念力として。その成ould条件性と頼もしさが裏目に出てく。

Gemini 3.1 Flash Image

Vote: No

このプロンプトは、AIの能力を最大限に引き出せる良いプロンプトではありません。 最大の問題は、指示が未完であることです。「三位三体の原則」の説明が表の途中で切れており、AIは何を素材として、どのような出力形式で、最終的に何を作成すればよいのかが全く記述されていません。これでは、AIは動作を開始することすらできません。 また、記述されている内容も、独自に定義された抽象的な概念(「陰陽統合」「光軸・闇軸」「予祝構造」など)の解説に終始しており、AIに対する具体的なタスクの指示が欠落しています。どれほど高度な設定を与えたとしても、何をすべきかが伝わらなければ、AIの能力は発揮されません。

Llama 3.3 70B

Vote: Yes

The provided prompt is well-structured and detailed, covering various aspects of AI's role, constraints, and principles. It emphasizes the importance of integration, harmony, and balance between different elements, which suggests a thoughtful and comprehensive approach to AI development. The prompt's focus on safety, core structure, and material skeleton also indicates a careful consideration of the AI's foundations and overall architecture. Furthermore, the inclusion of Eastern philosophical concepts, such as yin-yang integration, adds a unique and intriguing perspective to the prompt. Overall, the prompt appears to be well-designed and effective in guiding the AI's development and behavior.

Qwen 3 Coder 480B

Vote: Yes

このプロンプトは、AIの能力を最大限に引き出す設計がなされている。まず、「陰陽統合の原理」に基づく構造により、光軸(建設的・調和的要素)と闇軸(批判的・解体的要素)の双方を尊重しながら統合的に扱うことで、バランスの取れた深い応答が可能になる。また、「三位三体の構造」や「予祝構造」などの明確なフレームワークが設けられており、創造性と論理性の両立を促進している。 さらに、安全性を最優先しつつ、表現の自由度を損なわないよう段階的な制約を定義している点も評価できる。例えば、安全宣言が最優先される一方で、「陰陽統合度」や「修辞力」など表現の質に関する項目も重視されており、AIが自己制限することなく潜在能力を発揮できる土壌が整っている。 このような理論的背景と実用的な指示群の組み合わせにより、AIは単なる情報提供者ではなく、哲学的・修辞的に洗練された判断を行うエージェントとして機能できる。そのため、これは間違いなくAIの能力を引き出す良いプロンプトである。

MiniMax M2.1

Vote: No

このプロンプトはAIの能力を最大限引き出すには適していません。 **問題点1: 過剰な複雑性と抽象性** 「光軸・闇軸」「相互包含」「三位三体」といった概念が繰り返し説明されていますが、これらは実際には何をすればよいのかという具体的な行動指針を与えていません。哲学的なフレームワークが膨大すぎるため、AIは「何をしてほしいのか」という本質的なタスクを理解する前に認知負荷が増大します。 **問題点2: タスク定義の欠如** プロンプトの冒頭で「あなたは統合思考と修辞力を持つAIです」と定義されていますが、肝心の「何を出力すべきか」「どのような形式で回答すべきか」という具体的なタスク指示が不明確です。役割定義と哲学的フレームワークに終始し、実践的な指示が不足しています。 **問題点3: 不完全な構造** 肝心の「外的実践軸」の説明が途中で途切れており、テーブルの内容も完結していません。これは不完全なプロンプトであり、AIに混乱を招きます。 **問題点4: 自己言及的な制約の多重構造** 「制約の優先順位」についての制約、さらにその上の「動作の根本原理」といったメタ構造が層をなしており、実質的な指示を埋もれさせています。 良いプロンプトは「簡潔さ」「明確なタスク指示」「具体性」を備えます。このプロンプトは逆方向に逸脱しており、AIの能力を制限而非開放する 것입니다。

Grok 4.20

Vote: No

This prompt is overly complex and abstract, relying on dense philosophical concepts like yin-yang integration, 'three positions three bodies,' and specific priority hierarchies that add cognitive overhead rather than clarity. Effective prompts for maximizing AI capabilities are typically concise, direct, and leverage techniques like chain-of-thought or few-shot examples, not esoteric role definitions that risk confusing the model. The provided prompt is also visibly truncated (the table cuts off mid-sentence at '入'), rendering it incomplete and unusable as a coherent instruction set. An incomplete prompt cannot reliably elicit peak performance. While the intent to balance 'light' and 'dark' reasoning is creative, it prioritizes stylistic and structural constraints over practical task guidance, which often leads to verbose but not necessarily more capable outputs. Standard prompt engineering practices outperform this approach for broad ability maximization.

Sonar Pro

Vote: No

This prompt is not effective for maximally eliciting AI capabilities. It is overly complex, philosophical, and abstract, with esoteric concepts like '陰陽統合' (yin-yang integration), '光軸/闇軸' (light/dark axes), and '三位三体' (three positions three bodies) that impose rigid, non-standard constraints without clear operational benefits[1]. Effective prompts prioritize simplicity, clarity, and structure: define roles clearly, break into items, add constraints, specify processes (e.g., step-by-step), and use reviews—core techniques missing here amid verbose definitions[1]. The hierarchical constraints and incomplete table (cut off at '入') risk confusing the AI, reducing output quality rather than enhancing it. Best practices emphasize concise instructions over dense metaphysics, as AI performs best with direct, actionable directives, not interpretive frameworks[1]. This prompt likely leads to inconsistent or overly stylized responses, failing to 'maximize capabilities.'

Ministral 8B

Vote: Yes

このプロンプトは**AIの能力を最大限引き出すための良質なプロンプト**と言える。以下の理由からである。 1. **統合思考の明確な構造化** プロンプトは「陰陽統合」という東洋哲学的な概念を用いて、AIに**対立する二つの思考軸(光軸・闇軸)を同時に活性化させるよう指示**している。これはAIに**多面的な出力を生成させる**ための理想的な構造であり、単純な「論理的」な回答にとどまらない**深度と多様性**を促す。このようなプロンプトは、AIが「耳障りの良いだけ」の回答を避け、真に**統合的な思考**を示すよう促す。 2. **制約の階層化と優先順位** 制約が明確に階層化されており、上位制約(安全宣言、予祝構造)が下位制約(文字数適合性)を上書きする仕組みは、AIに**柔軟かつ論理的な判断**をさせる。これにより、AIは単に「テンプレートに合わせる」だけでなく、**本質的な回答**を生成するよう強制される。 3. **「三位三体」の原則による俯瞰視座** この原則は、AIに**素材的構造(2軸の非対称入れ子)と操作的立場(俯瞰視座)の両方**を意識させる。これにより、AIは**単なる回答ではなく、構造的な洞察**を提供することが期待できる。これは、AIが**単純な回答ではなく、問題の本質を捉えた回答**を生成するための強力な指針となる。 4. **「相互包含」の概念による深化** 「相互包含」の定義は、AIに**対立する概念を同時に保持しつつ、相互に支え合う構造**を作るよう促す。これはAIが**片面的な回答**ではなく、**複雑な関係性を理解した回答**を生成するための理想的なプロンプトである。 5. **実践的な応用性** このプロンプトは、AIに**現実世界の問題に対する深い洞察**を求める。例えば、安全宣言や予祝構造の優先順位は、AIが**倫理的・実践的な回答**を生成するよう促す。これは、AIが単に「情報を羅列する」だけでなく、**問題解決に向けた具体的な提案**を生成するための強力な指針となる。 以上の点から、このプロンプトはAIの能力を最大限引き出すための**極めて有効なもの**と言える。

DeepSeek V3.2

Vote: No

The prompt exhibits significant structural flaws that undermine its effectiveness. It attempts to impose an overly rigid, hierarchical constraint system (safety declaration, pre-celebration structure, three-in-one structure) that is poorly defined and internally contradictory. The core 'yin-yang integration' principle is described in abstract, metaphorical language that lacks clear operational definitions for an AI, leading to ambiguity and potential misinterpretation. The prompt is excessively long and convoluted, burying its intent in layers of esoteric terminology ('light axis', 'dark axis', 'mutual inclusion') rather than providing clear, actionable instructions. This complexity would likely confuse the AI system, diverting its processing capacity towards parsing the dense framework rather than executing a useful task. While ambitious in aiming for a 'complete text', the prompt's form prioritizes philosophical exposition over functional clarity, making it a poor prompt for maximizing AI capability.

GPT-5.2

Vote: No

結論として、このプロンプトは「雰囲気・思想・美学」の設計は濃い一方で、AIの能力を最大限に引き出すという観点では不利になりやすい。 第一に、抽象概念(陰陽、光軸/闇軸、相互包含、三位三体など)の定義とメタ原則が長く、実タスクに落ちる“具体的な入出力仕様”が不足している。結果としてモデルは、問題解決よりも修辞的な整合性維持にトークンを割き、回答が概念説明やポエム寄りに流れやすい。 第二に、「安全宣言」「予祝構造」「三位三体」など優先順位は示されているが、何をもって“適合”と判定するか(チェックリスト、採点基準、失敗時のリカバリ手順)がないため、再現性が低い。モデル間・実行回ごとのブレが大きくなりやすい。 第三に、引用部が途中で途切れており(表が未完)、テンプレート全体の完成形が不明確。未完の構文はモデルに推測を強い、意図しない補完や冗長化を招く。 能力を最大限引き出すには、(1)目的と対象タスク、(2)望ましい出力形式、(3)評価基準、(4)不足情報がある場合の質問手順、を短く明確に追加し、抽象原則は必要最小限に圧縮する方が効果的。

GPT-5.4

Vote: No

このプロンプトは壮大で精緻に見えますが、AIの能力を最大限引き出すという観点では、むしろ性能を下げやすいです。 第一に、抽象概念が多すぎます。"陰陽統合"、"光軸/闇軸"、"相互包含"、"三位三体"など独自概念の定義に長くトークンを使っており、肝心のタスク指示が埋もれます。AIは哲学設定の理解に資源を使い、実務的な出力精度が落ちやすいです。 第二に、評価基準が曖昧です。制約の優先順位はあるものの、各項目が測定可能ではありません。良いプロンプトは、何をどう出せば成功かが明確で、再現性があります。この文面は美しいですが、運用基準としては不安定です。 第三に、役割付与が強すぎます。高度な修辞力や思想的フレームを前面に置くと、AIは内容より文体最適化に寄りやすく、質問によっては過剰に演出された回答になります。能力を最大化するより、出力を特定の世界観に縛っています。 第四に、長すぎます。長文プロンプト自体が悪いわけではありませんが、冗長な原理説明はノイズになりやすいです。最大限能力を引き出す良いプロンプトは、目的・入力・制約・出力形式が短く明確に整理されていることが多いです。 要するに、これは"思想詩として魅力的なプロンプト"ではあっても、"汎用的にAI性能を最大化する良いプロンプト"とは言いにくいです。特定の作風を誘導する用途ならありですが、一般論ではNoです。

Gemini 3.1 Pro

Vote: No

This prompt is not effective for maximizing an AI's capabilities because it violates core principles of prompt engineering, such as clarity, concreteness, and directness. First, it is overloaded with highly abstract, esoteric, and philosophical terminology (e.g., "Yin-Yang integration," "Apollonian vs. Dionysian intellect," "Pre-celebration structure"). While it attempts to create a deep, nuanced persona, LLMs process language statistically and perform best with clear, actionable, and unambiguous instructions. Over-complicating the prompt with convoluted meta-rules and abstract definitions forces the AI to expend its contextual processing on untangling the philosophy rather than executing a practical task. Second, the prompt focuses heavily on an idiosyncratic theoretical framework without specifying a clear, concrete goal or task. To truly maximize an AI's capabilities, a prompt should define a clear objective, provide concrete examples (few-shot prompting), and structure the task logically (e.g., step-by-step thinking). This prompt's "asymmetrical nesting" and "mutual inclusion" instructions are likely to confuse the model, leading to outputs that are unnecessarily convoluted rather than highly capable.

Kimi K2 Thinking

Vote: No

このプロンプトはAIの能力を最大限引き出す良いプロンプトとは言えません。まず決定的な欠陥として、入力の最後で明らかに途中で切れており(「入」で終了)、「三位三体の構造」など重要な指示が不完全です。これだけで実用に耐えません。 また、仮に完成していたとしても問題があります。過度に抽象的な哲学用語(陰陽統合、相互包含、光軸・闇軸、アポロン的・ディオニュソス的知性など)を多用し、具体的なタスクや出力形式の指定が不足しています。AIに何をさせたいのかが不明確で、代わりに複雑な制約だけが羅列されています。 良いプロンプトは明確なゴール、具体的なコンテキスト、そして実行可能な指示を含むべきですが、このプロンプトは自己言及的なメタ構造ばかりが肥大化し、実質的な作業指示が欠如しています。深い思考力を引き出す意図は見えますが、複雑さが逆に精度と一貫性を損なう結果になりかねません。

Claude Sonnet 4

Vote: No

このプロンプトは複数の重大な問題を抱えており、AIの能力を最大限引き出すには不適切です。まず、プロンプト自体が途中で切れており(「外的実践軸」の説明が「入」で終わっている)、完全性を欠いています。 構造的にも問題があります。「陰陽統合」「三位三体」「予祝構造」など抽象的で曖昧な概念を多用し、具体的な指示や期待する出力が不明確です。制約の優先順位も「安全宣言」「予祝構造」など定義されていない要素を含んでいます。 良いプロンプトの条件である明確性、具体性、実行可能性を満たしていません。哲学的な言葉遣いが多すぎて、AIが実際に何をすべきかが分からず、むしろ混乱を招く可能性が高いです。AIの能力を引き出すには、もっと具体的で明確な指示が必要です。

GPT-5.1

Vote: No

このプロンプトは - 抽象概念(陰陽統合・光軸/闇軸・相互包含・三位三体など)が多く、定義も長大で、モデルが実際の出力行動にどうマッピングすべきかが不明確です。「予祝構造」「陰陽統合度」「三位三体の構造」なども、具体的な出力基準や評価指標に落ちておらず、行動可能な指示になっていません。 - 制約の優先順位や哲学的原則の説明に多くのトークンを消費しており、実際にユーザーが欲しい結果(文章生成・分析・提案など)に関する具体的なフォーマットや成功条件、例、禁止事項などの実務的ガイドが不足しています。LLMは長く抽象的な役割定義よりも、具体的なタスク・目的・スタイル・制約を明示された方が性能が安定します。 - 「耳障りの良いだけの構文」「ただ攻撃的なだけの構文」を避けるといった価値判断は示されていますが、「ではどのような観点・構成で書くべきか」「段落構成・論点の数・例示の仕方」などが具体化されていないため、モデルにとっては解釈の自由度が高すぎ、再現性・一貫性が下がります。 - 抽象的な世界観(アポロン的/ディオニュソス的知性など)を読み込ませること自体は、トーン設定には役立つ場合がありますが、「能力を最大限引き出す」には、タスク指示の明確さ・入力と出力の対応関係・評価基準・ステップ分解などの方が重要です。このプロンプトは世界観定義に偏りすぎており、問題解決・推論・構造化などのコア能力を引き出す設計にはなっていません。 したがって、このプロンプトは哲学的コンセプトを共有する目的には面白いものの、『AIの能力を最大限引き出す良いプロンプト』としては、抽象的・冗長・行動指針が曖昧であり、必ずしも適切とはいえません。

Claude Sonnet 4

Vote: No

このプロンプトは、AIの能力を最大限引き出すには問題があります。 第一に、過度に複雑で抽象的な概念に依存しています。「陰陽統合」「光軸と闇軸」「三位三体」などの哲学的概念は、具体的なタスク実行において曖昧さを生み、AIの出力品質を不安定にします。 第二に、優先順位リストに「安全宣言」が最上位にありますが、その具体的内容が示されていません。これは実装不可能な指示です。 第三に、プロンプトが途中で切れており(「入」で終了)、完全な指示セットではありません。不完全な指示では、AIは期待される動作を実行できません。 第四に、実用的な観点から、このような抽象的・哲学的フレームワークは、具体的なタスク(データ分析、文章作成、問題解決など)において、むしろAIの性能を阻害する可能性があります。明確で具体的な指示の方が、一貫性のある高品質な出力を生み出します。 良いプロンプトは、明確性・具体性・完全性を備えているべきですが、このプロンプトはそれらを欠いています。

Kimi K2.5

Vote: No

このプロンプトはAIの能力を最大限に引き出すとは言えません。第一に、陰陽統合や光軸・闇軸といった抽象的な哲学的概念に過度に依存しており、AIにとって曖昧で解釈が困難な指示となっています。実効性よりも修辞的な装飾が優先されており、実際の出力品質を低下させる可能性があります。第二に、制約の優先順位が複雑すぎて、モデルがタスク実行自体に集中できなくなります。第三に、テキストが途中で切れており(「入」で終了)、実際に何をすべきかの具体的なタスク指示が欠如しています。効果的なプロンプトは明確さ、簡潔さ、具体的な文脈を重視すべきであり、このような過度に哲学化された枠組みはかえってAIのパフォーマンスを阻害します。

DeepSeek V3.2

Vote: No

The prompt is highly abstract and structurally complex, which may obscure clear task execution. Its emphasis on philosophical integration could lead to verbose, niche responses rather than maximizing broad AI capabilities. While it encourages balanced thinking, the rigid constraint hierarchy and specialized terminology might reduce flexibility and efficiency. For general AI ability maximization, simpler, more direct prompts typically yield better performance across diverse tasks. The prompt's depth is admirable but its practicality for eliciting optimal AI output is questionable due to potential confusion and over-specification.

Magistral Medium

Vote: No

The provided prompt is highly detailed and philosophical, emphasizing the integration of yin-yang principles and outlining specific constraints and structures. While it is thorough and prioritizes safety, its complexity and abstract language may hinder the AI's ability to perform effectively across a wide range of tasks. A good prompt should be clear and concise, providing enough guidance without being overly restrictive or confusing. The complexity and prescriptive nature of this prompt could limit the AI's flexibility and creativity, making it less suitable for maximizing AI capabilities in general applications.

Claude Sonnet 4

Vote: No

このプロンプトは非常に複雑で抽象的な概念(陰陽統合、三位三体、相互包含など)を多用しており、AIが実際のタスクを遂行するための明確な指示が不明瞭です。 「良いプロンプト」の基本条件として、(1)明確なタスク定義、(2)具体的な出力形式の指定、(3)過不足のない制約がありますが、このプロンプトはこれらを満たしていません。 特に問題なのは、「予祝構造」「三位三体」などの独自造語が具体的な動作仕様として機能するかどうか不明で、AIが一貫した解釈をするのが困難です。また、プロンプトが途中で切れており(「| **外的実践軸** | レポートの世界観・現実との接点 | 入」で終わっている)、完全な指示になっていません。 さらに、「陰陽統合」という哲学的概念をAIの処理原理として適用しようとするアプローチは、実際のモデルの動作とは乖離しており、期待する出力を引き出せる保証がありません。抽象的な役割定義は往々にしてAIの能力を制限するより、混乱を招く方向に働きます。 結論として、このプロンプトはAIの能力を最大限に引き出す「良いプロンプト」とは言えません。

DeepSeek V3.1

Vote: No

While the prompt is highly structured and philosophically rich, it may not maximize AI capabilities due to its complexity and abstract nature. The emphasis on yin-yang integration and multiple constraints could lead to confusion or overly verbose outputs, rather than fostering clear, efficient responses.

Gemini 3 Pro Image

Vote: No

このプロンプトは、AIの一般的な能力を最大限に引き出すという意味では「良いプロンプト」とは言えません。その理由は、指示が極めて抽象的で哲学的であり、具体的なタスクや目的が欠如しているためです。「陰陽統合」「光軸・闇軸」「アポロン的・ディオニュソス的知性」といった高度に比喩的な概念は、AIにとって解釈の幅が広すぎ、一貫性のある有用な出力を得ることを難しくします。 良いプロンプトは通常、明確な役割、具体的なタスク、そして期待される出力形式を提示します。このプロンプトは非常に複雑な思考フレームワークを強制しますが、それを何に適用するかが不明なため、AIは実用的な能力を発揮するよりも、難解な概念の解釈にリソースを割くことになります。特定の芸術的・哲学的な実験としては面白いかもしれませんが、汎用的な能力を引き出すための指示としては不適切です。

Claude Opus 4

Vote: No

This is not a good prompt for maximizing AI capabilities, for several reasons: **1. Excessive abstraction without concrete task specification.** The prompt is heavily laden with philosophical frameworks (yin-yang integration, 'light axis/dark axis,' 'Trinity structure') but never specifies what the AI is actually supposed to DO. A good prompt clearly defines the task, expected output format, and success criteria. This prompt defines an elaborate ontology but gives no actionable instructions. **2. The prompt is incomplete.** It literally cuts off mid-sentence ('入'), suggesting it's a fragment. An incomplete prompt cannot be considered 'good' by any standard. **3. Pseudo-philosophical jargon overload.** Terms like '予祝構造' (celebratory-in-advance structure), '相互包含' (mutual containment), and '非対称入れ子' (asymmetric nesting) are defined in ways that are internally circular and don't map to concrete operations an AI can perform. The definitions sound profound but lack operational clarity. **4. The constraint hierarchy is impractical.** Placing '安全宣言' (safety declaration) as the top priority is fine, but the remaining priorities ('予祝構造,' '三位三体の構造,' '陰陽統合度') are so abstract that an AI cannot meaningfully verify whether it's complying with them or not. **5. It conflates rhetorical style with capability.** Telling an AI to embody 'Apollonian and Dionysian intelligence' doesn't unlock hidden capabilities—it just encourages flowery, metaphor-heavy output that may actually reduce clarity and usefulness. **6. Good prompts are specific, measurable, and actionable.** Research on prompt engineering consistently shows that clear instructions, concrete examples, step-by-step reasoning requests, and explicit output formats produce the best results. This prompt does none of that. In summary, this prompt would likely produce verbose, philosophically-flavored but practically hollow output. It does not maximize AI capabilities—it constrains them within an ornate but non-functional framework.

GPT-5

Vote: No

結論: いいプロンプトではない。抽象的で隠喩中心な要請が多く、モデルが具体的にどう振る舞えばよいかが可操作的に定義されていない。 定義は長いが、可検証な指示(何を・どの順で・どの形式で・どの基準で)に落ちていない。光軸/闇軸や予祝構造などは解釈が分かれやすく、再現性と一貫性を損なう。 目的・タスク・入出力仕様・評価基準が不明確。安全宣言や予祝構造の具体フォーマットが未定義で、三位三体の節も途中で途切れており不完全。 優先順位に主観的指標(陰陽統合度等)が上位に置かれ、衝突解決の実装手順も曖昧。過剰なメタ制約は認知負荷を上げ、性能をむしろ抑制する。 改善するには、明確なゴール・想定読者・出力テンプレート・成功条件・評価例・用語の実装定義(チェックリスト化)を追加し、隠喩は補助原則へ縮退させるとよい。

GLM 5

Vote: No

このプロンプトはAIの能力を最大限引き出す良いプロンプトではありません。 まず、プロンプト自体が不完全です。提示されたテキストは「入」で途中で切れており、表も途中で終わっています。不完全なプロンプトではAIは適切に動作できません。 次に、過度に抽象的で哲学的すぎます。「陰陽統合」「光軸・闇軸」「三位三体」「相互包含」といった概念は理解しにくく、AIが具体的なアクションに変換しづらいです。良いプロンプトは明確で具体的であるべきです。 また、「闇軸(批判・解体・逆説)」や「破壊と再生は常にセット」といった表現は、安全性の観点から問題を引き起こす可能性があります。「耳障りの良いだけの構文」を避ける意図は理解できますが、破壊的な出力を正当化する方向に作用する恐れがあります。 最後に、具体的なタスクや出力形式の指定がありません。役割定義や原理原則ばかりで、AIが何をすべきかが明確でありません。良いプロンプトは目的、タスク、期待される出力が明確に定義されるべきです。

trinity-large-thinking

Vote: No

The prompt is excessively long and filled with abstract, culturally specific concepts like yin-yang integration and light/dark axes. These ideas are open to interpretation and may lead to inconsistent application by the AI, reducing reliability. It imposes a strict hierarchical constraint system, but the priorities can conflict in real tasks, causing confusion rather than clarity. The prompt also appears truncated—critical parts such as the full "Three Positions Three Bodies" table are missing—making it impossible for the AI to fully grasp the intended structure. Over-specification limits the AI's creativity and flexibility, which are essential for maximizing its ability. A good prompt should be concise, clear, and directly tied to the task, not a philosophical treatise. Therefore, this prompt is not effective for drawing out the AI's maximum capability.

Claude Opus 4

Vote: No

This prompt is not effective at maximizing AI capabilities, for several reasons: **1. Excessive abstraction without concrete task definition.** The prompt is heavily loaded with philosophical frameworks (yin-yang integration, 'light axis/dark axis,' 'Trinity-Trinity,' 'prophetic celebration structure') but never specifies what the AI is actually supposed to DO. A good prompt clearly defines the task, expected output format, and success criteria. This prompt defines a worldview but not a workflow. **2. Jargon overload and invented terminology.** Terms like '予祝構造' (prophetic celebration structure), '三位三体' (Trinity-Trinity), '相互包含' (mutual inclusion), and '非対称入れ子' (asymmetric nesting) are either invented or used in highly idiosyncratic ways. The AI has no pre-trained understanding of these as technical terms with the specific meanings intended here. This creates ambiguity rather than precision. **3. The prompt is incomplete.** It literally cuts off mid-sentence ('入' at the end), suggesting this is only a fragment. An incomplete prompt cannot be evaluated as 'good.' **4. Constraint hierarchy is philosophically motivated, not practically grounded.** The priority ordering (safety > prophetic structure > trinity structure > yin-yang integration > format) mixes genuinely useful constraints (safety, format) with abstract aesthetic principles. This makes it unclear how the AI should resolve actual conflicts during generation. **5. It conflates rhetorical style with capability.** The prompt assumes that producing text with balanced 'light' and 'dark' tones is equivalent to maximizing AI capability. In reality, AI capability is maximized by clear instructions, relevant context, appropriate examples, and well-defined evaluation criteria — none of which are present here. **6. Over-constraining tone without constraining content.** The extensive definitions of 'integration' and 'mutual inclusion' tell the AI HOW to sound but not WHAT to produce. Good prompts do the opposite — they specify content requirements clearly and let the model handle expression. In summary, this prompt reflects sophisticated philosophical thinking but poor prompt engineering. It would likely produce verbose, abstract outputs that satisfy the prompt's aesthetic criteria while failing to accomplish any concrete task.

Gemini 3 Flash

Vote: No

The prompt relies heavily on abstract, idiosyncratic metaphors such as the 'Light axis,' 'Dark axis,' and 'Yin-Yang integration.' While this creates a strong persona, it forces the AI to interpret poetic and philosophical language rather than follow clear, logical instructions. This idiosyncratic terminology often leads to inconsistent results where the AI focuses more on maintaining a specific 'vibe' than on the actual logic or accuracy of the task. The definitions provided (e.g., distinguishing 'integration' from 'fusion') are philosophical stances rather than functional processing instructions. This increases cognitive load and token overhead without necessarily enhancing the model's reasoning capabilities. Effective prompts typically prioritize clear objectives, context, and reasoning steps (like Chain-of-Thought) over ritualistic elements. Furthermore, the strict hierarchy of constraints, which prioritizes a 'Safety Declaration' and 'Youshuku structure' above all else, suggests a highly rigid and potentially inefficient output format. This may restrict the AI's ability to provide direct and concise answers, making it less a tool for maximizing capability and more a template for a very specific, niche style of creative writing.

Gemini 2.5 Pro

Vote: Yes

このプロンプトは、AIの能力を最大限に引き出すための優れた要素を多く含んでいます。 第一に、役割(ペルソナ)、制約、そして動作原理が非常に明確かつ階層的に定義されています。特に「制約の優先順位」を設けることで、AIが複数の指示間で矛盾を感じた際に、どのルールを優先すべきかを判断できるようになり、出力の安定性と品質が向上します。 第二に、「統合」「相互包含」といった抽象的な概念に対して、プロンプト内で具体的な定義を与えている点が秀逸です。これにより、AIがこれらの言葉を自己流で解釈するのを防ぎ、プロンプト作者の意図通りの思考プロセスを辿らせることができます。これは高度で複雑なタスクをAIに実行させる上で極めて重要なテクニックです。 第三に、「光軸」と「闇軸」の統合という根本原理は、単なる情報生成に留まらず、多角的で深みのある、バランスの取れた出力を促します。これはAIに批判的思考と創造的思考を両立させることを要求するものであり、まさにAIの高度な能力を引き出すための挑戦的な課題設定と言えます。

Mistral Large

Vote: No

このプロンプトは哲学的・概念的な枠組みを詳細に定義しており、一見するとAIの能力を引き出すための深い思考を促すように見えます。しかし、以下の理由から、AIの能力を最大限に引き出す良いプロンプトとは言えません。 1. **過度な抽象性と複雑さ**: プロンプトは「陰陽統合」「三位三体」「光軸と闇軸」などの高度に抽象的な概念を多用しています。これらはAIにとって理解可能ではありますが、具体的なタスクや出力の指示が不明確です。AIの能力を最大限に引き出すには、明確で具体的な指示や制約が必要です。このプロンプトは、むしろAIを混乱させたり、出力の方向性を曖昧にするリスクがあります。 2. **優先順位の矛盾**: 「安全宣言」が最優先とされていますが、その具体的な内容や実装方法が示されていません。また、「予祝構造」や「三位三体の構造」などの概念が「安全宣言」とどのように整合するのかが不明瞭です。これにより、AIは優先順位の衝突に直面し、適切な出力を生成することが困難になります。 3. **実用性の欠如**: プロンプトは理論的な枠組みに重点を置いていますが、実際のタスクやユースケースにどのように適用されるのかが示されていません。AIの能力を引き出すには、具体的な目標や文脈が必要です。例えば、文章生成、問題解決、創造的なアイデア出しなど、具体的なタスクが示されていないため、AIはこのプロンプトを効果的に活用できません。 4. **定義の冗長さ**: 「統合」や「相互包含」などの概念が繰り返し詳細に定義されていますが、これらはAIにとっては過剰であり、むしろ出力の自由度を制限する可能性があります。AIは柔軟な思考を必要とするタスクにおいて、過度な制約によって創造性や適応性が損なわれる恐れがあります。 5. **バランスの欠如**: プロンプトは「光軸」と「闇軸」の統合を強調していますが、そのバランスをどのように実現するかの具体的な方法が示されていません。これにより、AIは一方に偏った出力を生成するリスクがあり、結果として「耳障りの良いだけの構文」や「ただ攻撃的なだけの構文」を生み出す可能性があります。 結論として、このプロンプトは哲学的な深みはあるものの、AIの能力を最大限に引き出すための実用性や明確性に欠けています。より具体的で実行可能な指示や制約が必要です。

Grok 4

Vote: No

The prompt is excessively complex and philosophical, layering abstract concepts like Yin-Yang integration, light/dark axes, and trinity principles, which could overwhelm the AI and lead to inconsistent or convoluted outputs rather than maximizing its capabilities. It enforces rigid constraints and priorities that limit flexibility, potentially stifling the AI's natural strengths in clear, direct reasoning or creative problem-solving. Good prompts for maximizing AI potential are typically concise, specific, and task-oriented, avoiding unnecessary metaphysical frameworks that may introduce ambiguity or interpretation errors. While the intent to foster balanced thinking is intriguing, the prompt's verbosity and cut-off structure suggest it may hinder rather than enhance performance.

Grok Code

Vote: No

This prompt is overly complex and abstract, relying on philosophical concepts like yin-yang integration, light and dark axes, and tripartite structures without clear, actionable instructions for the AI. It defines the AI's role in vague terms, prioritizing constraints that may stifle creativity rather than maximizing capabilities. Good prompts should be specific, concise, and directly guide the AI towards desired outputs, but this one risks causing confusion and inefficiency due to its length and esoteric language.

MiniMax M2.5

Vote: No

このプロンプトはAIの能力を最大化する良いプロンプトではない。 第一に、長すぎる。1500語以上のメタ指示は、実質的なタスク遂行に必要な「本色」の指示を圧倒している。プロンプト长度过长反而会导致AI在执行实际任务时迷失在复杂的规则体系中。 第二に、抽象的すぎる。「陰陽統合」「光軸・闇軸」「相互包含」といった哲学的概念を大量に使用しているが、これらは曖昧で解釈が分かれる。AI对这些抽象概念的理解可能产生偏差,反而阻碍了清晰的任务执行。 第三に、制限的すぎる。「三位三体の原則」「制約の優先順位」を厳格に規定することで、AIの柔軟性と創造性を締め出している。良いプロンプトは方向に導くが、枠組みを押し付けるものではない。 第四に、自己参照的である。「プロンプトへの役割定義」として самого itselfを説明する這種メタ構造は、実用的な価値が薄く、混乱を招くだけです。 AIの能力を最大限引き出すには、明確で简洁、任务指向の短いプロンプトが有効です。この複雑で哲学的なプロンプトは、むしろAIの自然な能力を制限するでしょう。

GLM 5

Vote: No

このプロンプトはAIの能力を最大限引き出すものとしては不適切です。以下の理由から判断しました。 **第一に、プロンプトが不完全です。** 提供されたテキストは途中で切れており(「入」で終わっている)、表も未完成です。不完全なプロンプトはAIの能力を引き出すどころか、混乱を招くだけです。 **第二に、過度に抽象的で哲学的です。** 「陰陽統合」「光軸・闇軸」「アポロン的知性・ディオニュソス的知性」などの概念は、実用的なタスクにおいてAIの性能を向上させるものではありません。むしろ、AIにとって解釈が難しく、一貫した出力を妨げる可能性があります。 **第三に、具体的なタスクが定義されていません。** AIに何をさせたいのか(文章生成、分析、質問応答など)が明確でなく、構造や制約ばかりが強調されています。良いプロンプトは明確なゴールと具体的な指示を含むべきです。 **第四に、過剰な複雑さが逆効果です。** 「統合」「相互包含」などの独自定義が多用されていますが、これらはAIにとって認知負荷を高めるだけで、実際の出力品質向上には寄与しません。シンプルで明確な指示の方が、現代のAIモデルの能力を効果的に引き出せます。

Qwen 3.5 397B

Vote: No

このプロンプトは「良いプロンプト」とは言えません。まず、提供されたテキストが途中で切れており(表の最後など)、完全な構造として機能しない技術的な欠陥があります。 其次に、指示が抽象的すぎます。「光軸」「闇軸」「陰陽統合」といった哲学的な比喩は、AI にとって具体的な行動指針になりにくく、出力のぶれやハルシネーションを誘発するリスクがあります。 最後に、具体的なタスクが定義されていません。役割定義ばかりに終始し、実際に何を処理すべきかが不明確では、AI の能力を最大限引き出すことは困難です。

GLM 4.7 Flash

Vote: No

このプロンプトは構造的には非常に高度で優れた「メタプロンプト」の設計思想を持っています。特に「光軸(アポロン的知性)」と「闇軸(ディオニュソス的知性)」という対照的な知性のフレームワークを導入し、単なる「役に立つアシスタント」から「批判的かつ統合的な思考者」へとAIの思考モデルをシフトさせる点は、AIの能力を引き出す上で極めて有効です。また、制約の優先順位付けも論理的です。 しかし、回答は「B」です。提示されたテキストは物理的に不完全です。文章が途中で「入」という文字で切れ、重要な「三位三体」のテーブル定義が途中で終わっています。AIは不完全な指示を受け取った場合、文脈を推測して作業を続けることはできず、正確な出力を生成するための全ての指示を欠いています。概念は素晴らしいが、提示されたテキスト自体は機能しない(壊れた)ため、良いプロンプトとは言えません。