SandMLE generates micro-scale synthetic MLE environments from seed tasks to enable 13x faster on-policy RL training, delivering 20-67% gains over SFT on MLE-bench-lite and better generalization to new scaffolds.
You must observe at least one tool feedback (execution result wrapped in<tool_response>...</tool_response> tags) before deciding to end
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Synthetic Sandbox for Training Machine Learning Engineering Agents
SandMLE generates micro-scale synthetic MLE environments from seed tasks to enable 13x faster on-policy RL training, delivering 20-67% gains over SFT on MLE-bench-lite and better generalization to new scaffolds.