提出DAGGER模型，让数学推理更抗干扰且省计算资源
DAGGER: Distractor-Aware Graph Generation for Executable Reasoning in Math Problems
- 将数学题解构为可执行的计算图，显式建模无关干扰项
- 在含干扰信息的孟加拉语数据集上，准确率比基线高14-20点，仅用11%的计算量
- 无需额外训练干扰样本，适合低资源、噪声环境下的数学推理
链式思维（CoT）提示广泛用于数学问题求解，包括低资源语言场景，但其在无关上下文下的表现仍缺乏系统研究。为此，我们构建了DISTRACTMATH-BN——一个在MGSM和MSVAMP基础上加入语义合理但无计算意义信息的孟加拉语基准。评估7个3B至12B参数模型发现：标准模型性能下降高达41分，而专用推理模型虽消耗五倍以上token，仍下降14至20分。我们提出DAGGER，将数学问题求解重构为可执行计算图生成，并显式建模干扰节点。通过监督微调与组相对策略优化对Gemma-3模型进行微调，在增强基准上达到与推理模型相当的加权准确率，但仅使用11%的计算量。重要的是，该鲁棒性无需在干扰样本上显式训练。结果表明，结构化中间表示相比自由形式方法，在噪声和低资源环境下更具鲁棒性和推理效率。
原文摘要 · Abstract (English)
Chain-of-Thought (CoT) prompting is widely adopted for mathematical problem solving, including in low-resource languages, yet its behavior under irrelevant context remains underexplored. To systematically study this challenge, we introduce DISTRACTMATH-BN, a Bangla benchmark that augments MGSM and MSVAMP with semantically coherent but computationally irrelevant information. Evaluating seven models ranging from 3B to 12B parameters, we observe substantial performance degradation under distractors: standard models drop by up to 41 points, while reasoning-specialized models decline by 14 to 20 points despite consuming five times more tokens. We propose †DAGGER, which reformulates mathematical problem solving as executable computational graph generation with explicit modeling of distractor nodes. Fine-tuning Gemma-3 models using supervised fine-tuning followed by Group Relative Policy Optimization achieves comparable weighted accuracy on augmented benchmarks while using 89 percent fewer tokens than reasoning models. Importantly, this robustness emerges without explicit training on distractor-augmented examples. Our results suggest that enforcing structured intermediate representations improves robustness and inference efficiency in mathematical reasoning compared to free-form approaches, particularly in noisy, low-resource settings.
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。