# Mind the Gaps: From Failure Attribution to Closed-Form Repair of Code Language Models arXiv: https://arxiv.org/abs/2610.05277 ยท HTML: https://mlsanity.com/paper/2610.05277/ ## Key facts - No citing papers in the map yet (data through 14 Oct 2026). - Builds on 49 of its 64 references that are in the map. - First posted on arXiv 4 Oct 2026. ## Builds on (most cited first, 30 of 49) - [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020) - [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021) - [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017) - [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022) - [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020) - [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022) - [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022) - [Towards Automated Circuit Discovery for Mechanistic Interpretability](https://mlsanity.com/paper/2304.14997/) (2023) - [MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategies](https://mlsanity.com/paper/2404.06395/) (2024) - [Eliciting Latent Predictions from Transformers with the Tuned Lens](https://mlsanity.com/paper/2303.08112/) (2023) - [Editing Factual Knowledge in Language Models](https://mlsanity.com/paper/2104.08164/) (2021) - [Transformer Feed-Forward Layers Build Predictions by Promoting Concepts in the Vocabulary Space](https://mlsanity.com/paper/2203.14680/) (2022) - [RepoCoder: Repository-Level Code Completion Through Iterative Retrieval and Generation](https://mlsanity.com/paper/2303.12570/) (2023) - [Towards Best Practices of Activation Patching in Language Models: Metrics and Methods](https://mlsanity.com/paper/2309.16042/) (2023) - [AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models](https://mlsanity.com/paper/2410.02355/) (2024) - [Transcoders Find Interpretable LLM Feature Circuits](https://mlsanity.com/paper/2406.11944/) (2024) - [Beyond Functional Correctness: Exploring Hallucinations in LLM-Generated Code](https://mlsanity.com/paper/2404.00971/) (2024) - [Model Editing Harms General Abilities of Large Language Models: Regularization to the Rescue](https://mlsanity.com/paper/2401.04700/) (2024) - [Interpreting Language Models with Contrastive Explanations](https://mlsanity.com/paper/2202.10419/) (2022) - [AnyEdit: Edit Any Knowledge Encoded in Language Models](https://mlsanity.com/paper/2502.05628/) (2025) - [Provable Repair of Deep Neural Networks](https://mlsanity.com/paper/2104.04413/) (2021) - [Knowledge Graph Enhanced Large Language Model Editing](https://mlsanity.com/paper/2402.13593/) (2024) - [LongCodeZip: Compress Long Context for Code Language Models](https://mlsanity.com/paper/2510.00446/) (2025) - [DeepLocalize: Fault Localization for Deep Neural Networks](https://mlsanity.com/paper/2103.03376/) (2021) - [LLMs Meet Library Evolution: Evaluating Deprecated API Usage in LLM-based Code Completion](https://mlsanity.com/paper/2406.09834/) (2024) - [DeepFault: Fault Localization for Deep Neural Networks](https://mlsanity.com/paper/1902.05974/) (2019) - [The Mirage of Model Editing: Revisiting Evaluation in the Wild](https://mlsanity.com/paper/2502.11177/) (2025) - [NNrepair: Constraint-based Repair of Neural Network Classifiers](https://mlsanity.com/paper/2103.12535/) (2021) - [CodeUpdateArena: Benchmarking Knowledge Editing on API Updates](https://mlsanity.com/paper/2407.06249/) (2024) - [Precise Localization of Memories: A Fine-grained Neuron-level Knowledge Editing Technique for LLMs](https://mlsanity.com/paper/2503.01090/) (2025) ## Abstract (arXiv) Code language models must be maintained like the software around them: when a library evolves, a model keeps writing the interface that it saw during training. Repairing the model itself lets one correction reach all downstream uses. Existing repair methods attribute a failure to neurons, select the highest-ranked ones, and apply a generic update. This pipeline assumes that the attributed neurons are the ones to patch and that a generic update fits every failure, and neither assumption has been examined. We examine both on executable API evolution tasks in Python and Rust with three code models and identify two gaps. The targeting gap separates the neurons that failure attribution targets from the neurons that can carry a patch: their top sets have a Jaccard overlap of only 0.15 to 0.20. The tailoring gap separates a generic update from a patch built for the failure: the patches that different carrier neurons need are nearly orthogonal. To address both gaps, we propose ASTRA. It targets neurons by contrastive semantics, an attribution that scores a neuron by its contribution to the logit contrast between the target token and the produced token. It then tailors the patch by solving one small linear system in closed form, which corrects all failing tokens of a sample jointly and needs neither an optimizer nor a backward pass. Contrastive semantics selects significantly better carrier neurons than gradient-based attributions in 3 of 6 settings and comparable ones in the others. On average, ASTRA reaches 66.7 percent Pass@1, against 48.4 percent for the best of AlphaEdit, STAR and low-rank adaptation, and repairs a sample in 2.7 seconds. It is the best method in all 6 settings and for every type of API change, and this advantage persists under an unseen phrasing of the test prompt. Its side effects on unrelated code are small on the large models and larger on the small one.