Researchers have developed a framework called TrojanGYM to detect and insert hardware Trojans in digital designs. The framework uses large language models (LLMs) to automatically propose and refine modifications that realize diverse triggers and payloads without impacting the design's functionality. This approach is designed to expose blind spots in existing detectors, which often overfit to specific patterns. The researchers claim that their method can increase detection rates from 0% to 60% relative to prior art on challenging benchmarks. --- Why it matters: This matters because hardware Trojans pose a significant threat to digital designs, and developing effective detection methods is crucial for ensuring security. TrojanGYM's ability to expose blind spots in existing detectors could lead to improved security measures and more robust detection techniques. Source: https://arxiv.org/abs/2601.17178