Skip to content
InfoResearchPreprintLLM-specific

Verified Netlist-to-RTL Recovery with LLM Agents

Published
Record updated
View JSON

Summary

LOCA (LOgic Cone Attack) is an LLM-based method that reverse engineers an anonymized gate-level netlist into RTL by splitting it into logic cones and having parallel agents rewrite them. Only RTL proven equivalent to the netlist is kept, and across 21 model-design pairs LOCA recovers 17, versus 6 with whole-netlist prompting. Per-run recovery across models rises from 5% to 74%, at a median cost below one dollar with a low-cost model.