iScript: A Domain-Adapted Large Language Model and Benchmark for Physical Design Tcl Script Generation

A collection of recent CS/SE arXiv papers introducing domain-adapted LLMs, benchmarks, datasets, and automation tooling focused on code generation, retrieval, verification, and robustness. Highlights include iScript — a Qwen3-8B adapted for Innovus Tcl script generation with a synthesized 10K (requirement, chain-of-thought, script) dataset and a two-step verification pipeline; CLARC — a C/C++ code-search benchmark stressing robustness (identifier anonymization, compilation to low-level targets); Vibe Code Bench — an autonomous browser-evaluated end-to-end web-app generation benchmark showing L