extending swe-2's reward function to steer the pareto frontier, Anish Lakkapragada

Experiments extending the reward function in Cognition’s SWE-2 model release.

saved by

• Akira Yoshiyama