erobinson/repro-toward-training-superintelligent-software-agents-through-self-play-swe-rl

Reproduction: Toward Training Superintelligent Software Agents through Self-Play SWE-RL

An open experiment logbook, published with Trackio.