RF-Agent: A Practical Framework for Building Language Agents for RFIC Design

2026-07-21

AI Summary

This study addresses the lack of domain-specific datasets and standardized benchmarks in radio-frequency (RF) integrated circuit design, which has hindered the effective application of large language models. To bridge this gap, the authors propose a textbook-driven, multi-agent QTSA knowledge distillation framework, leveraging seven canonical RF textbooks to construct the first reasoning dataset and multiple-choice evaluation benchmark for the field. The work systematically evaluates supervised fine-tuning (SFT) alongside three retrieval-augmented generation (RAG) strategies—semantic, keyword-based, and hybrid—demonstrating that domain-specific fine-tuning substantially enhances RF reasoning capabilities in small-to-medium-scale models. Among the RAG approaches, semantic retrieval yields the best performance, indicating that embedding alignment is particularly well-suited to the nuances of RF reasoning tasks.

Abstract

Large language models (LLMs) have driven rapid progress in electronic design automation (EDA), yet their application to radio-frequency (RF) circuit design remains limited by the scarcity of domain-specific datasets and standardized benchmarks. We present RF-Agent, which addresses this gap through textbook-driven knowledge distillation. A multi-agent Question-Thinking-Solution-Answer (QTSA) pipeline converts a subsection-level corpus from seven canonical RF textbooks into the first-of-its-kind RF-domain reasoning dataset (over 11,000 samples) with a dedicated multiple-choice benchmark. On this benchmark we study two adaptation strategies: supervised fine-tuning (SFT) and three retrieval-augmented generation (RAG) configurations (semantic, keyword, hybrid). Across multiple LLM families, domain-specific SFT significantly improves RF reasoning, especially for small and medium-sized models; among RAG configurations, semantic retrieval performs best, indicating embedding-based context alignment suits RF reasoning better than naive fusion. The dataset and benchmark provide a reusable foundation for future work on LLM-aided RF circuit design.

Problem

• RFIC design
• large language models
• domain-specific datasets
• standardized benchmarks
• electronic design automation

Innovation

• RF-Agent
• knowledge distillation
• reasoning dataset
• retrieval-augmented generation
• RFIC design