When Self-Evolution Backfires: Pre-Commit Gating against Skill Contamination in LLM Agents
It is shown the contamination is structurally irreversible: removing a source skill after the fact cannot erase the flawed reasoning its descendants have already inherited, so post-hoc rollback recovers only a small fraction of the lost performance, making skill admission a pre-commit necessity rather than a post-hoc fix.
Lin-Fang Shang, Ming Xu, Yi-Ding Sun et al.
· 1 citation