Table 1 Comparison of DR-CoT with chain-of-thought approaches.

Scratchpad⁷

Intermediate computations

Lacks refinement, inefficient context handling

Recursive reasoning for dynamic updates

Simple Prompting⁹

Step-by-step zero-shot prompts

Struggles with complex tasks

Dynamic context truncation, recursion

Plan-and-Solve^10,11

Task decomposition

Uses a single reasoning chain

Aggregates multiple chains via voting

Chain-of-Knowledge¹²

Integrates external data

Depends on external sources

Manages internal context efficiently

Decomposed Prompting^13,14

Modular breakdown

Weak error correction

Consensus-based voting reduces errors