BDH-CQ: In-Context Learning with Recurrent Latent Reasoning
from Daily Paper Cast (Test) Hosted by Jingwen Liang, Gengyu Wang
🤗 Upvotes: 235 | cs.NE, cs.AI, cs.LG, stat.ML 作者: Björn Engdahl, Adrian Kosowski, Jan Chorowski, Zuzanna Stamirowska, Przemysław Uznański, Junlin Jiang, Rohan Phadke, Remigiusz Kinas, Richard Zhong 标题: BDH-CQ: In-Context Learning with Recurrent Latent Reasoning Arxiv: http://arxiv.org/abs/2608.09888v1 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.
Embed this episode
No transcript yet.