Shengnan An

large language model (3) in-context learning (2) compositional generalization (2) few-shot learning (2) zero-shot learning (2) content moderation (1) empirical study (1) natural language understanding (1) instruction tuning (1) hierarchical reinforcement learning (1) nearest neighbor (1) context window (1) task adaptation (1) long context (1) long-context training (1) question-answer dataset (1) chain of thought reasoning (1) example selection (1) instruct tuning (1) context utilization (1)