A new multimodal benchmark of 3,614 circuit QA pairs shows that large language models perform worst on back-end layout and computation tasks, and that current models generally underperform on circuit design questions.
Verilogeval: Evaluating large language models for verilog code generation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
MMCircuitEval: A Comprehensive Multimodal Circuit-Focused Benchmark for Evaluating LLMs
A new multimodal benchmark of 3,614 circuit QA pairs shows that large language models perform worst on back-end layout and computation tasks, and that current models generally underperform on circuit design questions.