複雑なタスクを、より小さく相互に連結したプロンプトへと分解し、各出力を次のステップの入力として用いることで、構造化された推論パイプラインを構築する基本的なプロンプトエンジニアリング手法。多段階の問題における LLM の性能を大幅に向上させる。
Loading technique guide…
パターン: 複雑なタスクを、焦点を絞ったステップの直線的なパイプラインに分割する
理由: モデルの認知負荷を下げ、エラーを分離し、ツール統合とステップ単位のキャッシュを可能にする
重要なポイント: 構造化データ（JSON）による Output[N] → Input[N+1]。各ステップはキャッシュ可能
Discover how Sequential Chaining relates to other patterns
Prerequisites, next steps, and learning progression
No prerequisites needed - great for beginners!
Execute multiple chains concurrently for speed and multi-perspective analysis
💡 Natural next step to add concurrency to your linear chains
Parallel processing followed by aggregation - perfect evolution of sequential chains
💡 Scale your sequential logic to handle parallel data streams
Distribute sequential steps across multiple agents and gather results
💡 Transform linear chains into parallel distribution patterns
Fork sequential tasks into parallel subtasks and join when complete
💡 Add sophisticated parallel execution with synchronization points
DAG-based workflows with complex branching logic
💡 Advanced evolution for complex non-linear workflows
Single-prompt step-by-step reasoning without explicit chaining
💡 Simpler approach when you don't need separate API calls for each step
Interleaves reasoning and action steps with tool use
💡 Better when you need dynamic tool usage throughout the workflow
Multi-stage analysis and decision-making workflows using sequential chains
Sequential processing for document analysis and knowledge extraction
Code analysis and generation workflows
厳選された資料で理解を深めましょう
Chain-of-Thought Prompting Elicits Reasoning in Large Language Models (Wei et al., 2022)
Self-Consistency Improves Chain of Thought Reasoning (Wang et al., 2022)
Least-to-Most Prompting Enables Complex Reasoning (Zhou et al., 2022)
A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications (2024)