ATLAS：验证器引导的自适应潜在激活引导用于高效LLM推理
ATLAS: Verifier-Guided Adaptive Latent Activation Steering for Efficient LLM Reasoning
AI总结：
提出ATLAS框架，通过轻量级验证器动态调整推理时潜在状态引导策略，实现每步自适应控制，在数学和编码推理任务上提升准确率并减少测试时token使用。
摘要：
最近关于激活和潜在引导的研究表明，修改内部表示可以有效引导大型语言模型（LLMs）在不更新模型参数的情况下提高推理和效率。然而，大多数现有方法依赖固定引导策略和静态干预强度，这限制了它们在问题实例上的鲁棒性，并常常导致过度或不足引导。我们提出自适应测试时潜在引导（ATLAS），这是一个轻量级框架，通过训练好的、轻量级验证器在推理时动态控制引导决策。给定中间隐藏状态，验证器预测当前推理的质量，并自适应选择要应用的引导动作，实现每个示例和每个步骤的调整，且开销最小。ATLAS提供了一个统一框架，将学习到的潜在验证与测试时激活引导相结合，无需额外的LLM解码或推理时过程奖励模型调用即可实现自适应推理控制。在多个数学和编码推理基准上的实验表明，ATLAS始终优于普通解码和固定引导基线，在实现更高准确率的同时大幅减少测试时token使用。这些结果表明，验证器引导的潜在适应提供了一种有效且可扩展的机制，可以在不牺牲解决方案质量的情况下控制推理效率。所有源代码将公开提供。
英文摘要：
Recent work on activation and latent steering has demonstrated that modifying internal representations can effectively guide large language models (LLMs) toward improved reasoning and efficiency without updating model parameters. However, most existing approaches rely on fixed steering policies and static intervention strengths, which limit their robustness across problem instances and often result in over- or under-steering. We propose Adaptive Test-time Latent Steering (ATLAS), a lightweight framework that dynamically controls steering decisions at inference time using a trained, lightweight verifier over the latent states. Given intermediate hidden states, the verifier predicts the quality of ongoing reasoning and adaptively selects which steering action to apply, enabling per-example and per-step adjustment with minimal overhead. ATLAS provides a unified framework for combining learned latent verification with test-time activation steering, enabling adaptive reasoning control without additional LLM decoding or inference-time process reward model calls. Experiments on multiple mathematical and coding reasoning benchmarks show that ATLAS consistently outperforms both vanilla decoding and fixed steering baselines, achieving higher accuracy while substantially reducing test-time token usage. These results demonstrate that verifier-guided latent adaptation provides an effective and scalable mechanism for controlling reasoning efficiency without sacrificing solution quality. All source code will be publicly available.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
