LLM4Cov: Execution-Aware Agentic Learning for High-coverage Testbench Generation

ArXi:2602.16953v3 Announce Type: replace Execution-aware LLM agents offer a promising paradigm for learning from tool feedback, but such feedback can be expensive and slow to obtain, making online reinforcement learning (RL) less practical in certain scenarios. High-coverage hardware verification exemplifies this challenge due to its reliance on industrial simulators and non-differentiable execution signals. We propose LLM4Co, an offline agent-learning framework that models verification as single-step state transitions guided by deterministic evaluators. Building on this formulation, we.