aiqbench analysis
Spaghetti Architect introduces a novel framework for generating controlled code datasets by systematically introducing redundancy and anti-patterns from a clean intermediate representation to test model robustness against data contamination and difficulty scaling. This approach provides researchers with labeled benchmarks that decouple intrinsic problem complexity from incidental presentation noise, offering a rigorous method to evaluate how AI models handle artificially degraded or "messy" code across multiple languages.
Spaghetti Architect: A Contamination-Resistant, By-Construction-Labelled, Multi-Language Code Dataset Generator
arXiv:2607.18642v1 Announce Type: new Abstract: Mined code corpora are abundant but uncontrolled: a snippet's semantics, surface "messiness," and difficulty are whatever the wild contained; there is no known-optimal reference to grade against; and any public sample may already sit in a model's training set. We present Spaghetti Architect, a tool that mints code datasets with the control such corpora lack. An anti-optimization transpiler maps a clean, language-agnostic JSON intermediate representation to deliberately redundant, fully-flattened programs in five languages (Python, JavaScript, Go, Java, C++); every program is compiled, run, and checked against a reference oracle, so each instance is correct by construction. The clean IR is a known-optimal reference, messiness is dialed by strictly-nested anti-pattern profiles, each instance is labelled along two orthogonal difficulty axes, intrinsic (problem size) and incidental (presentation at fixed semantics), and contamination is resisted by minting fresh variants from a private held-out seed. We give construct-validity evidence that the quality order moves established complexity and readability metrics, and report baselines on a four-model open ladder: exact match rises with scale, and the intrinsic knob collapses arithmetic-aggregation accuracy of even the strongest model to zero. Further, development-set scores equal freshly re-minted held-out counterparts within $|\Delta|\le 0.012$ (comprehension) and $\le 0.011$ (refactoring); on identical programs, refactoring equivalence ($0.73 \rightarrow 0.99$) is scale-invariant while output prediction collapses; and ablating the generator's self-annotations shows they inflate the weakest model an order of magnitude more than the strongest ($-0.173$ vs $-0.017$): the annotated ladder resolves one of three adjacent pairs where the unannotated resolves all three. Open source (MIT), dependency-free, archived under a persistent DOI.
Abstract:Mined code corpora are abundant but uncontrolled: a snippet's semantics, surface "messiness," and difficulty are whatever the wild contained; there is no known-optimal reference to grade against; and any public sample may already sit in a model's training set. We present Spaghetti Architect, a tool that mints code datasets with the control such corpora lack. An anti-optimization transpiler maps a clean, language-agnostic JSON intermediate representation to deliberately redundant, fully-flattened programs in five languages (Python, JavaScript, Go, Java, C++); every program is compiled, run, and checked against a reference oracle, so each instance is correct by construction. The clean IR is a known-optimal reference, messiness is dialed by strictly-nested anti-pattern profiles, each instance is labelled along two orthogonal difficulty axes, intrinsic (problem size) and incidental (presentation at fixed semantics), and contamination is resisted by minting fresh variants from a private held-out seed. We give construct-validity evidence that the quality order moves established complexity and readability metrics, and report baselines on a four-model open ladder: exact match rises with scale, and the intrinsic knob collapses arithmetic-aggregation accuracy of even the strongest model to zero. Further, development-set scores equal freshly re-minted held-out counterparts within $|\Delta|\le 0.012$ (comprehension) and $\le 0.011$ (refactoring); on identical programs, refactoring equivalence ($0.73 \rightarrow 0.99$) is scale-invariant while output prediction collapses; and ablating the generator's self-annotations shows they inflate the weakest model an order of magnitude more than the strongest ($-0.173$ vs $-0.017$): the annotated ladder resolves one of three adjacent pairs where the unannotated resolves all three. Open source (MIT), dependency-free, archived under a persistent DOI.
32 pages. Under review at the Journal of Data-centric Machine Learning Research (DMLR). Code: this https URL (artifact archived at doi:https://doi.org/10.5281/zenodo.21033174)
Machine Learning (cs.LG); Software Engineering (cs.SE)
arXiv:2607.18642 [cs.LG]
(or arXiv:2607.18642v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.18642 arXiv-issued DOI via DataCite (pending registration)
Full text under the publisher's open licence · 300 words
Related coverage
Article: Governing AI in the Cloud: A Practical Guide for Architects
InfoQ · Jun 15, 2026
LiveLaw Supreme Court Half-Yearly Digest: January - June 2026 [Part 2]
livelaw.in · Sep 13, 2026
The Continual Constitutional Convention: How language models revise the documents that govern them
AI Alignment Forum · Oct 13, 2026
Technical Manual for a Toolkit for Measuring Contextual Individuation in Transformer Language Models
arXiv · Sep 7, 2026