🧠 Optimized for DeepSeek-R1 & DeepSeek-V3 Reasoning
DeepSeek-R1 Reasoning Architect
DeepSeek-R1 excels at pure reasoning, mathematics, and formal auditing. Build structured prompts that enforce deep internal reflection, backtracking on errors, and rigorous multi-step verification.
Reasoning Architecture Parameters
Target Thinking Token Depth: Exhaustive (~4,000 tokens)
Direct CoT Exhaustive Reflection Formal Proof Level
Generated DeepSeek-R1 System Prompt
🚀 Tested with DeepSeek-R1 671B MoE