让小模型更易接续，提升推理能力
FutureBridge: Token Selection Beyond Local Preference in Collaborative Decoding
- 基于未来推理潜力筛选候选词，而非仅依赖大模型偏好
- 在数学推理上使小模型准确率提升35.1%相对效果
- 适合需要高效协作推理的轻量级模型应用
Token级协作使大语言模型（LLM）在预测与小语言模型（SLM）分歧时能提供帮助。现有方法要么使用LLM生成的干预词，要么用LLM的下一个词概率对候选词排序，均依赖于LLM的局部偏好，而忽略了某些由LLM选择的词可能难以被SLM后续利用。我们提出FutureBridge，根据候选词对SLM后续推理的支持程度，对联合的LLM-SLM候选词进行排序。训练时，由经过答案验证的LLM轨迹提供固定共享未来，冻结的SLM在此共同上下文中评估每个候选词。由此产生的反事实得分监督一个仅观察当前状态和候选词的轻量级重排序器。推理时，FutureBridge仅用LLM扩展候选池，选出一个词后即返回给SLM，不生成或追加未来后缀。在五个数学推理基准上，FutureBridge将Qwen3-1.7B SLM的Math Avg提升35.1%相对效果。结果表明，通过建模候选词是否可被接收的SLM用于继续推理，能显著优于仅依赖大模型局部偏好的方式。
原文摘要 · Abstract (English)
Token-level collaboration allows a large language model (LLM) to assist a small language model (SLM) when their predictions diverge. Existing methods either use LLM-generated intervention tokens or rank candidates with the LLM's next-token probabilities. Both rely on the LLM's local preference, even though an LLM-selected token may be difficult for the SLM to build on. We present FutureBridge, which ranks joint LLM-SLM token candidates according to how well they support the SLM's subsequent reasoning. During training, an answer-verified LLM trajectory supplies a fixed shared future, and a frozen SLM evaluates every candidate under this common context. The resulting counterfactual scores supervise a lightweight token reranker that observes only the current state and candidate token. At inference, FutureBridge uses the LLM only to expand the candidate pool, selects one token, and returns generation to the SLM without generating or appending a future suffix. Across five mathematical reasoning benchmarks, FutureBridge improves the Qwen3-1.7B SLM's Math Avg. by 35.1% relative to greedy SLM decoding. These results indicate that token selection benefits from modeling whether the receiving SLM can use each candidate to continue reasoning, rather than relying on the LLM's local preference alone.
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。