通过优化答案边缘分布，实现更快更强的自一致性推理
Self-Consistency via Marginal Sharpening
- 将自一致性从后验投票转为推理时目标，聚焦答案支持路径数量
- 在数学与编码任务上超越标准幂采样，速度提升数个量级
- 纯自回归并行采样算法，无需额外训练，适配高阶推理场景
推理时采样可在不增加训练的情况下激发语言模型的强大推理能力。现有幂采样方法通过锐化完整生成结果的分布来实现，偏好模型认为单独可能的完成。我们指出，这并非推理应关注的对象：一个完成将推理过程与最终答案纠缠在一起，真正重要的是答案是否被多个合理推理路径支持。因此，我们把目标从完整输出分布转向锐化的答案边缘分布，使自一致性成为推理时的目标而非事后投票。令人惊讶的是，这一边缘目标可高效近似：我们提出一种简单、纯自回归的并行采样算法，近似从锐化答案边缘采样，在数学和编码基准上表现优于标准幂采样，且速度提升数个量级。
原文摘要 · Abstract (English)
Inference-time sampling can elicit strong reasoning abilities from language models without additional training. Existing power-sampling methods do so by sharpening the distribution over full generated outputs, favoring completions that are individually likely under the model. We argue that this is the wrong object to target for reasoning: a completion entangles a reasoning trace with a final answer, whereas what matters is whether an answer is supported by many plausible reasoning paths. We therefore shift the target from the full-output distribution to the sharpened answer marginal, making self-consistency an inference-time objective rather than a post-hoc voting criterion. Surprisingly, this marginal target admits an efficient approximation: we propose a simple, purely autoregressive parallel sampling algorithm that approximately samples from the sharpened answer marginal, eliciting stronger performance than standard power sampling on mathematics and coding benchmarks while being orders of magnitude faster.
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。