AI RESEARCH
Toward Scalable Verifiable Reward: Proxy State-Based Evaluation for Multi-turn Tool-Calling LLM Agents
arXiv CS.AI
•
ArXi:2602.16246v3 Announce Type: replace Interactive large language model (LLM) agents operating via multi-turn dialogue and multi-step tool calling are increasingly used in production. Benchmarks for these agents must both reliably compare models and yield on-policy