共享词元预算让思维链反噬准确率，长思考反而挤掉答案
The Coupling Tax: How Shared Token Budgets Undermine Visible Chain-of-Thought Under Fixed Output Limits
- 思维链与答案共用输出词元预算，导致长推理挤占答案空间
- 在GSM8K和MATH-500上，不思考模式在2048词元内始终优于或等于思考模式
- 提出分预算生成策略，使准确率提升至83.6%，适合资源受限的推理部署
思维链推理常被视为通过延长思考时间提升模型准确率的方法。本文发现一种反向效应：当推理过程与最终答案共享输出词元预算时，过长的推理轨迹反而会压缩答案空间。在Qwen3三个规模模型上测试GSM8K、MATH-500及五个BIG-Bench Hard任务，结果表明，在2048词元以内，非思考模式在GSM8K和MATH-500上始终表现不差或更优；更难任务则需更大预算才出现交叉点。我们提出截断-浪费分解模型：Acc_think(b) = α_c F_L(b) + α_t (1 - F_L(b))，可基于推理长度与准确率统计预测该交叉点，并解释了Qwen家族中的逆缩放现象。对DeepSeek-R1-Distill-Llama-8B的复现也呈现相同规律。为缓解此问题，采用分预算生成策略，全量MATH-500上IRIS达到74.0%准确率，增强提取变体达78.8%，固定非正例门控方案达83.6%。结果表明，测试时推理应视为预算分配问题，而非仅关注推理长度。
原文摘要 · Abstract (English)
Chain-of-thought reasoning is often treated as a monotone way to improve language-model accuracy by letting a model think longer. We identify a countervailing effect, the coupling tax: when reasoning traces and final answers share one output-token budget, long traces can crowd out the answer they are meant to support. Across GSM8K, MATH-500, and five BIG-Bench Hard tasks with Qwen3 models at three scales, non-thinking mode matches or outperforms thinking mode on GSM8K and MATH-500 at every budget up to 2048 tokens, while harder tasks shift the crossover to larger budgets. We derive a truncation-waste decomposition, $\mathrm{Acc}_{\mathrm{think}}(b)=α_c F_L(b)+α_t(1-F_L(b))$, that predicts this crossover from chain-length and accuracy statistics and explains inverse scaling within the Qwen family. A DeepSeek-R1-Distill-Llama-8B replication shows the same pattern under a different thinking interface. As a mitigation, split-budget generation decouples reasoning and answer budgets; on full MATH-500, IRIS reaches 74.0% accuracy, a strengthened extraction variant reaches 78.8%, and a fixed non-oracle SC+IRIS gate reaches 83.6%. The results show that test-time reasoning should be evaluated as a budget-allocation problem, not only as a question of whether longer traces are available.
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。