让大模型像人一样思考前先评估，提升推理效率与准确率
Before you <think>, monitor: Implementing Flavell's metacognitive framework in LLMs
- 引入认知心理学框架，分三阶段迭代监控、生成、验证
- GSM8K上准确率达75.42%，比同类方法高8%以上，尝试次数少35%
- 适合需要高质量推理且关注过程可控性的研究者
当前增强大模型推理的方法主要分为两类：一类是监控-生成型（如Plan-and-Solve、SELF-DISCOVER），擅长策略规划但缺乏对策略有效性的验证机制；另一类是生成-验证型（如Self-Verification、SELF-REFINE），可迭代优化输出但缺乏任务评估即盲目生成。二者分离导致效率低下——策略失败无反馈，优化缺乏战略基础。本文引入Flavell的认知监控模型（1979），构建三阶段迭代系统，实现监控-生成-验证的闭环。在GSM8K数据集上，初步结果显示准确率为75.42%，优于SELF-REFINE（68.44%）和Self-Verification（67.07%），且平均尝试次数从2.0降至1.3，推理成本增加27%-37%。结果表明，前置监控可生成更高质量初始解，减少后续修正需求，但其泛化能力仍需在更广泛任务中验证。
原文摘要 · Abstract (English)
Current approaches to enhancing LLM reasoning follows two isolated paradigms: Monitor-Generate methods like Plan-and-Solve (Wang et al., 2023) and SELF-DISCOVER (Zhou et al., 2024) excel at strategic planning but lack mechanisms to verify whether selected strategies succeed; while Generate-Verify approaches like Self-Verification (Weng et al., 2022) and SELF-REFINE (Madaan et al., 2023) iteratively refine outputs but commence generation blindly without task assessment. This separation creates inefficiencies -- strategies fail without feedback, and refinement occurs without strategic grounding. We address this gap by implementing Flavell's cognitive monitoring model (1979) from the broader Monitor-Generate-Verify framework (Oh and Gobet, 2025), operationalising it as a three-phase iterative system. On GSM8K, preliminary results show 75.42% accuracy versus 68.44% for SELF-REFINE and 67.07% for Self-Verification, while requiring fewer attempts (1.3 vs 2.0) at 27-37% increased inference cost. These initial findings suggest upfront monitoring produces higher-quality initial solutions that reduce refinement needs, though evaluation beyond arithmetic reasoning is needed to establish generalisability.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。