Paper Overview
Field: Chip/Hardware Verification Author: FNU Aditi Published: 2026-09-22 arXiv: 2609.26751
Summary
Large language models are increasingly used to generate SystemVerilog Assertions from natural-language specifications and register-transfer-level designs. Existing datasets and benchmarks support important goals such as large-scale training, formal evaluation, specification-to-assertion generation, and mutation-based testing. A complementary need is to study whether a generated assertion captures externally observable behavior or depends on incidental details of one RTL implementation.
EquivSVA is a formally verified dataset organized around behavior families. Each family contains four structurally distinct RTL implementations of the same externally observable behavior, shared interface-level gold properties, three controlled mutants, and formal-validation evidence.
Dataset Contents
- 120 behavior families across 12 categories
- 480 reference RTL implementations
- 914 gold properties
- 360 mutants
- Fixed family-safe train, development, and test splits
- Of 293 interface-only generated properties, 93 are formally sound
- The number of sound properties varies across equivalent implementations for 14 of 24 test families
Every final family passes a fixed 17-job validation suite covering:
1. RTL equivalence 2. Gold-property proofs 3. Property reachability 4. Mutant distinguishability 5. Gold-property checks on mutants
Case Study
As a small demonstration of the analyses enabled by the dataset, the authors evaluate the publicly released, Apache-2.0-licensed Qwen2.5-Coder-7B-Instruct model on the held-out test split:
Resources
The dataset, generators, validation scripts, and case-study artifacts are publicly released at:
https://github.com/aditigupta96/EquivSVA