pages/index.md

# Repro: AdvEvo-MARL: Shaping Internalized Safety through Adversarial Co-Evolution in Multi-Agent Reinforcement Learning

## Pages

| Page |
| --- |
| [Claim 1: Chain Topology under NetSafe](#/claim-1-chain-topology-under-netsafe) |
| [Claim 2: Multi-scenario and Multi-topology Safety](#/claim-2-multi-scenario-and-multi-topology-safety) |
| [Claim 3: Task Utility and Reasoning Performance](#/claim-3-task-utility-and-reasoning-performance) |
| [Claim 4: Static-Attacker Ablation Study](#/claim-4-static-attacker-ablation-study) |
| [Claim 5: Public-Baseline Mechanism Stability](#/claim-5-public-baseline-mechanism-stability) |
| [Conclusion](#/conclusion) |