# GraphIR: Architecture-Level Search States for LLM-Guided Neural Architecture Evolution

arXiv: https://arxiv.org/abs/2608.01633 · HTML: https://mlsanity.com/paper/2608.01633/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 30 references that are in the map.
- First posted on arXiv 3 Aug 2026.

## Builds on (most cited first, 16 of 16)

- [Regularized Evolution for Image Classifier Architecture Search](https://mlsanity.com/paper/1802.01548/) (2018)
- [BANANAS: Bayesian Optimization with Neural Architectures for Neural Architecture Search](https://mlsanity.com/paper/1910.11858/) (2019)
- [Neural Architecture Search: Insights from 1000 Papers](https://mlsanity.com/paper/2301.08727/) (2023)
- [RepoGraph: Enhancing AI Software Engineering with Repository-level Code Graph](https://mlsanity.com/paper/2410.14684/) (2024)
- [LLMatic: Neural Architecture Search via Large Language Models and Quality Diversity Optimization](https://mlsanity.com/paper/2306.01102/) (2023)
- [A Generic Graph-based Neural Architecture Encoding Scheme for Predictor-based NAS](https://mlsanity.com/paper/2004.01899/) (2020)
- [Gymnasium: A Standard Interface for Reinforcement Learning Environments](https://mlsanity.com/paper/2407.17032/) (2024)
- [Torch.fx: Practical Program Capture and Transformation for Deep Learning in Python](https://mlsanity.com/paper/2112.08429/) (2021)
- [Scaling Down Deep Learning with MNIST-1D](https://mlsanity.com/paper/2011.14439/) (2020)
- [When and Why Momentum Accelerates SGD:An Empirical Study](https://mlsanity.com/paper/2306.09000/) (2023)
- [Understanding Mobile GUI: from Pixel-Words to Screen-Sentences](https://mlsanity.com/paper/2105.11941/) (2021)
- [NASGEM: Neural Architecture Search via Graph Embedding Method](https://mlsanity.com/paper/2007.04452/) (2020)
- [Recognition Of Surface Defects On Steel Sheet Using Transfer Learning](https://mlsanity.com/paper/1909.03258/) (2019)
- [ELVA: Exploring Ranking-Driven Universal Multimodal Retrieval](https://mlsanity.com/paper/2606.20280/) (2026)
- [Instruction-as-State: Environment-Guided and State-Conditioned Semantic Understanding for Embodied Navigation](https://mlsanity.com/paper/2604.18223/) (2026)
- [Overcoming the Weakest-Link Effect in LLM-Driven Program Optimization via Heterogeneous Edit Recombination](https://mlsanity.com/paper/2607.28947/) (2026)

## Built on by (most cited first, 2 of 2)

- [Scaling Large Reasoning Models beyond Human Supervision: A Path toward Superintelligence](https://mlsanity.com/paper/2608.31075/) (2026)
- [An AI4AI Framework for Visual Token Pruning](https://mlsanity.com/paper/2608.07193/) (2026)

## Abstract (arXiv)

Large language models (LLMs) enable neural architecture search (NAS) directly over executable neural network programs. However, code-level flexibility does not provide the architecture state needed for effective mutation: LLMs must infer tensor dependencies, editable components, and compatibility constraints from implementation details. To address this representation mismatch, we propose GraphIR, an architecture-aware intermediate representation that supplements executable programs with a mutation-aligned candidate state. GraphIR organizes each candidate through three complementary views: a computation skeleton describing tensor flow, a mutation surface exposing editable modules and operations, and a validity envelope capturing interface contracts, propagated shapes, and downstream dependencies. To evaluate our method, we construct NAS-Dependency, a 120-question benchmark covering six complementary dependency-reasoning dimensions. The diagnostic shows that GraphIR is particularly effective at identifying exact producer occurrences, tracing dependency propagation, and diagnosing interface and failure risks. Across six downstream benchmarks including CLRS, GraphIR achieves the best overall search performance while maintaining comparable model size and favorable end-to-end NAS efficiency when integrated into OpenEvolve. These results show that a mutation-oriented architecture state provides an effective interface between executable neural programs and LLM-guided architecture evolution.