out-of-domain generalization
This describes a model's ability to perform well on data that comes from different distributions or domains than the training data. Enhancing out-of-domain generalization is a challenge in building robust AI systems.
- Delving into RL for Image Generation with CoT: A Study on DPO vs. GRPO
- Distilling LLM Agent into Small Models with Retrieval and Code Tools
- Enigmata: Scaling Logical Reasoning in Large Language Models with Synthetic Verifiable Puzzles
- Omni-R1: Reinforcement Learning for Omnimodal Reasoning via Two-System Collaboration
- Retro-R1: LLM-based Agentic Retrosynthesis
- Towards A Generalist Code Embedding Model Based On Massive Data Synthesis
- True Zero-Shot Inference of Dynamical Systems Preserving Long-Term Statistics