Explore-Execute Chain: Towards an Efficient Structured Reasoning Paradigm

2025-09-28

AI Summary

Existing LLM inference architectures—such as Chain-of-Thought—conflate high-level planning with low-level execution, resulting in inefficiency, limited path exploration, and poor interpretability. To address this, we propose the Explore-Execute Chain (E²C), the first framework to decouple reasoning into two distinct phases: *exploration* (generating abstract, high-level plans) and *execution* (deterministically unfolding those plans). This design enables efficient test-time scaling and cross-domain transfer while reducing computational overhead and enhancing controllability over reasoning paths. Methodologically, E²C integrates supervised fine-tuning (SFT) with a novel data synthesis strategy and incorporates reinforcement learning to optimize plan fidelity and execution determinism. On AIME’2024, E²C achieves 58.1% accuracy using fewer than 10% of the decoding tokens required by baseline methods. Furthermore, EF-SFT—a variant optimized for medical tasks—outperforms standard SFT by 14.5% accuracy using only 3.5% more tokens, establishing a new state-of-the-art.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Chain-of-Thought (CoT) and its variants have markedly advanced the reasoning abilities of Large Language Models (LLMs), yet their monolithic and auto-regressive architecture inherently conflates high-level strategic planning with low-level step-by-step execution, leading to computational inefficiency, limited exploration of reasoning paths, and reduced interpretability. To overcome these issues, we propose the Explore-Execute Chain ($E^2C$), a structured reasoning framework that decouples reasoning into two distinct phases: an exploratory phase that stochastically generates succinct high-level plans, followed by an execution phase that deterministically carries out the chosen plan. Our approach incorporates a two-stage training methodology, which combines Supervised Fine-Tuning (SFT) - augmented by a novel data generation algorithm enforcing strict plan adherence - with a subsequent Reinforcement Learning (RL) stage that capitalizes on the informativeness of exploration and reinforces the determinism of execution.This decomposition enables an efficient test-time scaling strategy: on AIME'2024, $E^2C$ Test Time Scaling reaches 58.1% accuracy using <10% of the decoding tokens required by comparable methods (e.g., Forest-of-Thought), sharply cutting self-consistency overhead. For cross-domain adaptation, our Exploration-Focused SFT (EF-SFT) fine-tunes with only 3.5% of the tokens used by standard SFT yet yields up to 14.5% higher accuracy than standard SFT on medical benchmarks, delivering state-of-the-art performance, strong generalization, and greater interpretability by separating planning from execution. The code and pre-trained models for the project are available at: https://github.com/yks23/Explore-Execute-Chain.git

Problem

• Decouples reasoning into exploration and execution phases
• Reduces computational overhead while improving accuracy
• Enables efficient cross-domain adaptation with minimal fine-tuning

Innovation

• Decouples reasoning into exploratory and execution phases
• Uses two-stage training with SFT and reinforcement learning
• Achieves high accuracy with reduced decoding tokens