通过抑制中间层表征坍缩，显著提升Transformer的数学推理能力。
Seq-VCR: Preventing Collapse in Intermediate Transformer Representations for Enhanced Reasoning
- 引入序贯方差-协方差正则化，增强中间层表征多样性。
- 在5×5乘法任务中达99.5%准确率，远超同规模模型和GPT-4。
- 无需思维链标注，适合强化基础推理能力的模型改进。
Decoder-only Transformers在复杂推理任务中表现不佳，尤其在需要多步操作的算术推理方面。本文发现模型中间层表征坍缩是限制其推理能力的关键因素。为此，提出序列方差-协方差正则化（Seq-VCR），通过提升中间表示熵值来防止坍缩。结合虚拟暂停标记替代思维链标记，该方法显著提升算术推理性能。在具有挑战性的5×5整数乘法任务中，本方法达到99.5%精确匹配准确率，大幅优于同规模模型（0%）和使用五次提示的GPT-4（44%）。在算术表达式和最长递增子序列（LIS）数据集上也取得更优结果。研究强调防止中间层表征坍缩对增强Transformer推理能力的重要性，并证明Seq-VCR是一种无需显式思维链监督的有效解决方案。
原文摘要 · Abstract (English)
Decoder-only Transformers often struggle with complex reasoning tasks, particularly arithmetic reasoning requiring multiple sequential operations. In this work, we identify representation collapse in the model's intermediate layers as a key factor limiting their reasoning capabilities. To address this, we propose Sequential Variance-Covariance Regularization (Seq-VCR), which enhances the entropy of intermediate representations and prevents collapse. Combined with dummy pause tokens as substitutes for chain-of-thought (CoT) tokens, our method significantly improves performance in arithmetic reasoning problems. In the challenging $5 \times 5$ integer multiplication task, our approach achieves $99.5\%$ exact match accuracy, outperforming models of the same size (which yield $0\%$ accuracy) and GPT-4 with five-shot CoT prompting ($44\%$). We also demonstrate superior results on arithmetic expression and longest increasing subsequence (LIS) datasets. Our findings highlight the importance of preventing intermediate layer representation collapse to enhance the reasoning capabilities of Transformers and show that Seq-VCR offers an effective solution without requiring explicit CoT supervision.
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。