SPADE: Self-Play in Adaptive Synthetic Executable Environments Paper • 2608.19197 • Published 1 day ago • 38
SPADE Checkpoints Collection The six SPADE models: Qwen3 4B, 8B, and 30B-A3B, each trained on games and on tool use. • 6 items • Updated about 11 hours ago
SPADE Grounding Corpora Collection The source documents the Environment Designer builds from, 15k per setting, one per prompt. • 2 items • Updated about 11 hours ago
SPADE Synthetic Environments Collection The environments SPADE generated during training, and the GPT-5.5 pools used as the fixed baseline. • 4 items • Updated about 10 hours ago