BDH-CQ：结合循环潜态推理的上下文学习
BDH-CQ: In-Context Learning with Recurrent Latent Reasoning
浏览论文内容
中文总结 AI 辅助
该研究提出BDH-CQ模型，结合上下文学习与循环潜态推理，在ARC-AGI-1评估集上实现高成本效率，突破了该基准的成本-准确率帕累托前沿。
中文摘要 AI 辅助
我们提出了BDH-CQ，一种结合上下文学习与循环潜态推理的推理模型。推理时输入会持续更新模型的循环记忆；模型通过在高维潜态空间中迭代计算来解决查询，无需将中间推理过程用语言表述。我们在公开的ARC-AGI-1评估集上对该模型进行评估，并使用受控的类ARC干预措施，研究它从演示中学习到什么、应用推断出的变换的一致性如何，以及哪些概念仍然较难。1.5亿参数的配置在每个任务计算成本为0.0007美元时，达到了29.5%的pass@2指标。该工作点突破了此前报道的ARC-AGI-1成本-准确率帕累托前沿，在基准成本效率方面建立了新的技术水平。
英文摘要
We introduce BDH-CQ, a reasoning model that combines in-context learning with recurrent latent reasoning. Inputs presented at inference time continuously update the model's recurrent memory; the model then solves a query through iterative computation in a high-dimensional latent space, without verbalizing its intermediate reasoning. We evaluate the model on the public ARC-AGI-1 evaluation set and use controlled ARC-like interventions to study what it learns from demonstrations, how consistently it applies an inferred transformation, and which concepts remain difficult. A 150M-parameter configuration reaches 29.5% pass@2 at a computed inference cost of \$0.0007 per task. This operating point breaks through the previously reported ARC-AGI-1 cost-accuracy Pareto frontier, establishing a new state of the art in benchmark cost efficiency.
发表机构
- Pathway
- Bielik AI
- New York University（纽约大学）
机构由 AI 辅助整理，请以论文原文为准。
