Dataset Hub

ARC (Abstraction and Reasoning Corpus) is a dataset jointly created by Gwangju Institute of Science and Technology and Korea University, intended to evaluate and improve AI systems' abstract reasoning capabilities. It contains various complex grid‑editing tasks with large action spaces and diverse task types. The dataset was generated using the Gymnasium environment by defining specific action and state spaces to simulate ARC challenges. ARC is primarily used in reinforcement learning to develop and test AI models capable of solving complex reasoning problems.

PC‑Gym is an open‑source toolkit developed by Imperial College London for evaluating reinforcement‑learning algorithms on chemical process‑control problems. The dataset includes multiple simulated chemical‑process environments featuring nonlinear dynamics, disturbances and constraints. It provides custom constraint handling, disturbance generation and reward‑function design to accelerate research at the intersection of machine learning and process systems engineering. By offering a standardized platform, PC‑Gym bridges theoretical RL advances with real industrial process‑control applications, giving researchers a tool to explore data‑driven control solutions.

--- license: apache-2.0 dataset_info: features: - name: prompt dtype: string - name: summary_1 dtype: string - name: summary_2 dtype: string splits: - name: train num_bytes: 1697095 num_examples: 1000 download_size: 906302 dataset_size: 1697095 ---

The V‑D4RL dataset provides pixel‑based analogues of the D4RL benchmark tasks derived from the dm_control suite and extends two state‑of‑the‑art online pixel‑based continuous control algorithms, DrQ‑v2 and DreamerV2, to offline settings. It includes data of varying difficulty across multiple environments such as walker_walk, cheetah_run, and humanoid_walk, along with corresponding benchmarks and algorithm evaluations.