Concept
AEWM-RFT
A method for reinforcement learning fine-tuning based on AEWM-corrected trajectories, enabling the agent to internalize the corrected reasoning capability into model parameters without invoking AEWM at runtime.
Timeline (last 90 days)
Oct 15
AEWM-RFT shows that corrected trajectories can be written into the agent's own capabilities, and afterwards, without calling AEWM, the agent still outperforms the version trained only on its own successful trajectories by 2.2, 2.6 and 2.5 percentage points.
Unverified50%