English(EN) Unlocking Token Rewards via Training-Free Reward Attribution

新方法在无需重新训练的情况下从 LLM 中提取代币级奖励

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-15 04:00

研究人员开发了一种新颖的无训练方法，可以从现有的深度奖励模型中提取代币级奖励信号。该方法通过估计单个代币的影响来将其与整体过程奖励相关联，这种影响是通过用语义上为空的代币替换代币时奖励的变化来计算的。通过采用基于梯度的估计器和一阶泰勒近似，该方法仅需要一次前向和后向传播，从而显著提高了计算效率。在推理基准上的实验表明，策略优化和 LLM 推理泛化能力得到了显著提升，在 MathVista 和 AIME24 上的性能也有显著提高。 AI

影响该技术可以通过在无需额外的奖励模型训练的情况下进行精确的信用分配，从而增强 LLM 的推理能力和策略优化效率。

排序理由该集群包含一篇学术论文，详细介绍了 LLM 奖励归因的新方法。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 14 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇学术论文，详细介绍了 LLM 奖励归因的新方法。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, model release; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Sitong Wu, Haoru Tan, Bin Xia, Xichen Zhang, Jingyao Li, Shaofeng Zhang, Xiaojuan Qi, Bei Yu, Jiaya Jia · 2026-10-15 04:00

通过无训练奖励归因解锁代币奖励

arXiv:2610.18704v1 Announce Type: new Abstract: In this paper, we propose an extremely efficient, training-free method to extract token-level reward signals directly from an existing deep reward model. Our core idea is to attribute the overall process reward to individual tokens …

本文如何被排名

报道来源 [1]

通过无训练奖励归因解锁代币奖励

相关实体

相关话题