Information-Gain Rewards over Diversity-Pruned Tests: GT-Anchored Verifier Co-Training for Reliable Code Generation
arXiv:2609.21208v1 Announce Type: new Abstract: Self-play methods that co-train a single language model as both coder and test author promise to move code-generation RL beyond fixed test suites, but they suffer from two coupled pathologies: permissiveness collapse, where pass-rate rewards are maxim…
Read the full story at arXiv cs.AI ↗
Timeline · 1 report
- 2026-09-22 04:00 · arXiv cs.AI
Information-Gain Rewards over Diversity-Pruned Tests: GT-Anchored Verifier Co-Training for Reliable Code Generation