LegoFlow datasets
AI & ML interests
We study code intelligence.
Recent Activity
Papers
LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents
SWE-Review: Closing the Loop on Issue Resolution with Agentic Code Review
SWE-Review: Closing the Loop on Issue Resolution with Agentic Code Review. Benchmark, trajectories, training data, and models for agentic code review.
Harness-native RL for coding agents: the trained policy and the training task index.
-
LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents
Paper • 2608.17393 • Published • 27 -
Lego-X/qwen3_5_35b_a3b_ohsdk_200k_rl
Text Generation • 36B • Updated • 112 • 4 -
Lego-X/qwen3_5_35b_a3b_cc_200k_rl
Text Generation • 36B • Updated • 72 • 1 -
Lego-X/qwen3_5_35b_a3b_oc_200k_rl
Text Generation • 36B • Updated • 70 • 1
LegoFlow datasets
Harness-native RL for coding agents: the trained policy and the training task index.
-
LEGO-RL: Harness-Native Reinforcement Learning for Coding Agents
Paper • 2608.17393 • Published • 27 -
Lego-X/qwen3_5_35b_a3b_ohsdk_200k_rl
Text Generation • 36B • Updated • 112 • 4 -
Lego-X/qwen3_5_35b_a3b_cc_200k_rl
Text Generation • 36B • Updated • 72 • 1 -
Lego-X/qwen3_5_35b_a3b_oc_200k_rl
Text Generation • 36B • Updated • 70 • 1
SWE-Review: Closing the Loop on Issue Resolution with Agentic Code Review. Benchmark, trajectories, training data, and models for agentic code review.