reasoning-degeneration-dev/researchgym-continual-learning-gpt4o-20260311

researchgym-continual-learning-gpt4o-20260311: Harbor dataset on Hugging Face. ResearchGym agent traces (1 runs). Tasks: tasks/test/continual-learning. Models: openai/gpt-4o. Converted from inspect ai eval logs to ATIF/Harbor format.