Benchmarking General Purpose In-Context Learning
通用人工智能的奇思妙想试验场
Laboratory Toward General Intelligences
通用上下文学习及其训练和评测
(Benchmarking General Purpose In-Context Learning).
信息瓶颈限制的神经网络可塑性规则进化
(Evolving Plastisity with Information Bottleneck).
基于随机世界中大规模元训练的上下文强化学习
(OmniRL:Large-Scale Meta-Training in Randomized Worlds).