04:00
2026-09-24
arxiv.org
ai-agents
TwinCheck: Evidence-Grounded Negative-Twin Verification for Stateful Tool Agents
TwinCheck, an inference-time verification policy for stateful tool agents, raised task success for GPT-5.6 Sol from 45.3% to 58.5% on 159 multi-turn BFCL V4 tasks with complete exact-replay pairs, acc…