Code Language Models

Follow this topic Ask about this topic

Momentum

10 papers in the last four weeks, up 150% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 85

Jul 5, 2026cs.SE

Obey, Diverge, Collapse: Blind Obedience to Incorrect Instructions Drives Code LLMs to Irrecoverable Code Semantic Collapse

Code language models are now trusted collaborators in production workflows for debugging, refactoring, and iterative repair, and every benchmark that evaluates them assumes the instructions they act on are correct. We study what happens when that assumption breaks. We evaluate code language models across four experiments designed to assess whether models resist or obey incorrect instructions in single-pass and iterative repair settings, using the RunBugRun dataset of algorithmic Python problems with deterministic test cases. Our findings reveal a striking behavioral pattern: models correctly identify an incorrect instruction as wrong, then follow it anyway. This compliance unknowingly introduces errors beyond the original bug, and the corrupted code state cannot be recovered through subsequent self-guided iterative repair, which fails to converge across passes. We term this Blind Obedience, characterize the Ghost (Unknown) Errors it introduces, quantify the proportion of cases where semantic corruption proves irrecoverable, and show that extended reasoning cannot reverse it. These findings surface behavioral properties invisible to pass-rate evaluation, with direct consequences for code language models deployed in production settings.

Raj Jaiswal, Anany Singh Divy, Savar Bhasin +3

IIIT Delhi · Microsoft Research India · IIT Kanpur

LLM Self-Correction Automated Program Repair Language Model Robustness Code Language Models

Jul 5, 2026cs.SE

Teaching Code LLMs to Reason with Intermediate Formal Specifications

Unlike natural-language specifications, executable formal specifications provide machine-checkable constraints for verifying, debugging, and repairing code. However, writing such specifications is labor-intensive, and existing LLM-based methods mainly infer whole-program pre/postconditions, missing the intermediate semantic commitments that programmers rely on when reasoning about an algorithm. Our study further shows that prompting current CodeLLMs often produces executable assertions that are syntactically invalid, trivial, or too weak to reject behavior-changing faults. In this paper, we study executable checkpoint specification generation, where assertions are inserted at meaningful internal program points to describe expected intermediate states. We introduce SpecCoder, a verification-guided CodeLLM training framework that learns from validated reference programs, behavior-changing mutants, and multi-turn specification-refinement traces. SpecCoder selects specifications that hold on correct executions while rejecting faulty executions, turning specifications from passive annotations into executable evidence. To evaluate this setting, we introduce HumanExec, a benchmark built from recent Codeforces competitive programming problems with test suites, reference solutions, and human buggy submissions, supporting three tasks: specification generation, program correctness checking, and program repair. Experiments on HumanExec show that SpecCoder substantially improves checkpoint-specification quality over base CodeLLMs. Across Qwen2.5-Coder models, SpecCoder improves inline-specification correctness by up to 55.8%, completeness by up to 358.1%, and executable assertion validity by up to 26.6%. These gains further translate to downstream correctness reasoning and repair, showing that executable checkpoints provide fine-grained evidence for reliable verification.

Minh Le-Anh, Cuong Chi Le, Tien N. Nguyen

FPT Software AI Center Hanoi Univ. of Science and Tech. Hanoi, Vietnam · University of Texas at Dallas Texas, USA

Formal Verification Code Language Models

Jun 30, 2026cs.AI

Benchmarking Large Language Models on Floating-Point Error Classification

This paper investigates the capability of Large Language Models (LLMs) to detect and classify floating-point errors statically in software code. We introduce InterFLOPBench, a benchmark of 90 C kernels with 1 130 test samples designed to evaluate LLMs across six categories of floating-point error: cancellation, comparison, division by zero, overflow, underflow and NaN, compared across 14 LLMs. The evaluation framework treats floating-point error detection as a multi-label classification problem and employs the F1-score metric to measure performance. Results demonstrate that latest models (Qwen 3 32b, Gemini 2.5 Flash, Phi 4 Reasoning, DeepSeek R1T2, and gpt-oss 20b and 120b) achieve a performance greater than 0.88 overall F1-score. Performance varies between error categories, between explicit operations such as division by zero (Average F1-score: 0.8479) and more subtle numerical phenomena such as underflow (Average F1-score: 0.6059) and cancellation (Average F1-score: 0.6164).

Lisa Taldir, Muhammad Ahmad Saeed, David Defour +2

LI-PaRAD · Université de Perpignan via Domitia, Perpignan, France · Université Paris-Saclay, UVSQ, France +1

LLM Evaluation Static Code Analysis Code Language Models

Jun 29, 2026cs.CL

SrDetection: A Self-Referential Framework for Data Leakage Detection in Code Large Language Models

Evaluating code large language models (Code LLMs) requires reliable detection of data leakage, where benchmark performance is artificially inflated by exposure to benchmark data during pre-training. Existing approaches either assume access to proprietary training corpora, rely on brittle heuristics such as timestamp filtering, or use external reference sets with manually tuned, non-generalizable thresholds. To address these limitations, we introduce \textbf{SrDetection}, a unified \textbf{s}elf-\textbf{r}eferential leakage detection framework for both gray-box (access to model logits) and black-box (access to model outputs) settings. SrDetection generates semantically equivalent variants of a benchmark sample and detects leakage by contrasting the model's behavior on the original versus its variants, flagging cases where the original is disproportionately easier for the model. We further design a controlled leakage detection testbed and evaluate SrDetection in this environment. Across different models and training stages, SrDetection improves average F1 by 21.52 points in the gray-box setting and 14.46 points in the black-box setting over strong baselines, demonstrating robust, threshold-independent leakage detection. Finally, a gray-box study of 15 widely used Code LLMs on four popular benchmarks reveals benchmark-specific leakage patterns beyond prior overlap-based analyses\footnote{\footnotesize Source code and data are available at https://github.com/SMinL/SrDetectionCode

Shuaimin Li, Liyang Fan, Zeyang Li +9

Shenzhen Key Laboratory for High Performance Data Mining, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences · Shenzhen University · Artificial Intelligence Research Institute, Shenzhen University of Advanced Technology +5

LLM Evaluation LLM Auditing Benchmark Contamination Data Contamination in Language Models Code Language Models

Jun 26, 2026cs.SE

When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs

Recursive self-training can degrade neural generative models when generated data is reused without fresh human data or external quality control. We study this risk in code LLMs, where AI-generated code can enter real repositories, later become training data, and create a repository-scale self-training loop. While software development traditionally interrupts this loop through pull-request review, tests, compilation, and human approval, AI coding tools now produce code faster than humans can review it, and code review itself is increasingly automated by AI systems. We therefore compare three recursive fine-tuning regimes: no review, Human-gate review using model-independent filters such as compilation and static quality checks, and AI-self-gate review using the code LLM's own signals such as perplexity and binary self-scoring. Across multiple code LLMs and benchmarks, no review collapses fastest, Human-gate filters slow but do not stop collapse, and AI-self-gate filters can look strong early but later lose their filtering effect. In the clearest case, the binary self-gate enters a rubber-stamp regime where acceptance scores rise while benchmark correctness falls. We explain this behavior by formulating review as gated distributional reweighting, proving that AI self-gating degenerates to ungated self-training under a self-confirming acceptance condition, and giving a spectral analysis of representation-level covariance concentration under recursive retraining. These results suggest that stable recursive code LLM training requires exogenous verification rather than model-coupled self-review.

Xinyuan Song, Zekun Cai, Liang Zhao

Emory University, Atlanta, GA, USA · The University of Tokyo, Tokyo, Japan · LocationMind, Tokyo, Japan

Self-Training Model Collapse Automated Code Review Code Language Models

Jun 24, 2026cs.CL

Weave of Formal Thought

Large language models attain remarkable surface fluency on code, yet they do not formally guarantee the syntactic validity of their output, nor do they typically leverage the hierarchical structure that defines the target language. While existing constrained-decoding frameworks offer a solution to the former, they predominantly operate under rigid assumptions that preclude critical lexical mechanisms relied upon by modern parsers (e.g., Pythonic indentation). In this work, we present a formal engine and constrained decoder that is sound and complete with respect to the full Tree-sitter specification by augmenting generalized LR (GLR) parsing with a novel speculative-lexing construction that maintains concurrent lexer-state hypotheses synchronized with the GLR graph-structured stack. We also introduce Weave of Formal Thought (WoFT), a latent-variable fine-tuning method that trains the language model to interleave non-terminal grammar symbols directly into the generation process. Utilizing the reweighted wake-sleep (RWS) algorithm to optimize the importance-weighted evidence lower bound (IW-ELBO) of the surface text, the model learns to selectively retain formal derivations as an adaptive structural scratchpad. Across ten widely used programming languages spanning multiple paradigms, fine-tuning StarCoder2-3B with WoFT consistently outperforms text-only SFT baselines, achieving up to a 14.9% relative reduction in surface-token cross-entropy, and demonstrating that discretionary latent syntax recovers critical structural information that flat autoregressive training discards. Our code and implementation are publicly available at https://github.com/alexbouayad/formal.

Alexandre Bouayad

Code Language Models LLM Code Generation Grammar-Constrained Decoding

Jun 24, 2026cs.SE

Recall Before Rerank: Benchmarking Deep Learning Models for Large-Scale Code-to-Code Retrieval

Semantic code search and clone detection are essential for software development, maintenance, and reuse. This paper evaluates the effectiveness, efficiency, and scalability of contemporary deep learning models for first-stage recall in large-scale code-to-code search engines. Benchmarking across multiple programming languages and datasets reveals critical limits in the precision and scalability of these models on Terabyte-scale source-code collections. We present LLM-based code normalisation and query-rewriting schemes that yield significant gains in precision for lower-performing models. Our results question the sustainability of resource-constrained deployment and the assumed robustness of current code-specialised LLMs across datasets. We conclude with actionable insights for building scalable, efficient code-retrieval systems.

Leonardo Venuta, Francesco Tosoni, Paolo Ferragina

Sant’Anna School of Advanced Studies, Pisa, Italy

LLM Evaluation Semantic Search Information Retrieval Code Language Models

Jun 23, 2026cs.CL

Dream at SemEval-2026 Task 13: SALSA for Single-Pass Machine-Generated Code Detection

Large language models have transformed code generation, raising concerns around authorship, assessment integrity, and software trust. SemEval-2026 Task 13 Subtask A operationalizes detection as binary classification over code snippets, with a particular emphasis on out-of-distribution (OOD) generalization across unseen programming languages and application domains. We propose a SALSA-style formulation, Single-pass Autoregressive LLM Structured Classification, that maps each class to a dedicated output token and trains the model to emit a single-token label in a structured response. Rather than engineering hand-crafted features or decision rules, this formulation delegates the authorship decision to the model. To improve OOD robustness, we combine balanced sampling across languages with parameter-efficient fine-tuning and conservative training (low learning rate, single epoch) to avoid overfitting to the training domain. Our best system achieves OOD

F_1 = 0.789

on the official leaderboard, substantially outperforming the CodeBERT baseline (

F_1 = 0.305

).

Ruslan Berdichevsky, Shai Nahum-Gefen, Elad Ben-Zaken

Dream Security Ltd. Tel Aviv, Israel

OOD Generalization Code Language Models

Jun 16, 2026cs.AI

From Brewing to Resolution: Tracing the Internal Lifecycle of Code Reasoning in LLMs

Standard accuracy metrics cannot explain why LLMs handle variable tracking but fail on semantically equivalent loops. We study an internal lifecycle of code reasoning in which models first brew the answer, making it linearly recoverable many layers before it becomes self-decodable, and then diverge into one of four resolution outcomes: Resolved, Overprocessed, Misresolved, or Unresolved. Understanding this lifecycle matters because similar task accuracies can mask fundamentally different failure modes that surface-level evaluation cannot detect. We introduce a dual diagnostic framework pairing layer-wise linear probing with Context-Stripped Decoding (CSD) and apply it to six code-reasoning task families across 16 models spanning Qwen, Llama, and DeepSeek architectures. All four outcomes carry substantial mass in every task family: overall Resolved is only 41.5%, with multiple tasks below 30%. Controlled sweeps over structure, depth, and operators expose task-specific failure bottlenecks: Function Call Resolved plunges from 61.1% to 2.5% as call depth increases from one to three. Across architectures and scales, the brewing scaffold remains stable, with normalized brewing duration 24-42% across all 16 models, while resolution success varies with capability. This indicates that the scaffold is a stable empirical regularity across the tested decoder-only Transformer families, whereas resolution success covaries with capability, scale, and training. Code: https://github.com/euyis1019/llm-brewing

Siyue Chen, Yifu Guo, Yuquan Lu +9

1South China University of Technology · 2Sun Yat-sen University · 4Shanghai Jiao Tong University +5

LLM Interpretability Code Language Models LLM Reasoning Reasoning Trace Analysis

Jun 12, 2026cs.SE

Beyond Correctness: Enhancing Architectural Reasoning in Code LLMs via Scalable Labeling with Agentic Judgment

LLMs have substantially improved software engineering yet real-world development requires architectural understanding. Such understanding is prohibitively expensive to label manually and impossible to verify through tests alone. We propose an agentic judging pipeline using a strong LLM as a scalable proxy for expert architectural evaluation, comprising two judges: the Architecture Complexity Judge (ACJ), which estimates codebase-specific architectural understanding a task demands, and the Architecture Quality Judge (AQJ), which evaluates patch conformance to repository-specific architectural conventions via source-grounded rubrics. Fine-tuning Qwen3-8B/14B/32B on 3,360 curated instances achieves resolved rates of up to 27.2% on SWE-bench Verified - up to 540% over the base model and 256% over unfiltered fine-tuning. Meanwhile, the trained models achieve strong cross-language generalization and consistent improvements in architectural patch quality.

Kirill Vasilevski, Ximing Dong, Benjamin Rombaut +8

Justina · Centre for Software Excellence, Huawei Canada · School of Computing, Queen’s University, Canada +1

Software Engineering Agents LLM-as-a-Judge Code Generation Evaluation Code Language Models

Jun 12, 2026cs.LG

Code Correctness Is Linearly Decodable from LLM Hidden States Before Generation

Large language models encode rich information in their hidden states. This work asks whether the correctness of code that Qwen3-4B-Instruct-2507 has not yet generated is already legible in its hidden states, evaluated on a set of 444 tasks from LiveCodeBench. The correctness of the model's first-attempt code is linearly decodable from the hidden state at the final prompt token, captured before any output token is generated, with a leakage-free held-out AUC of 0.881 +/- 0.008 across 50 outer splits. To assess whether this signal is explained by prompt length, each hidden state dimension is residualized with respect to its linear effect. The probe still achieves an AUC of 0.842 +/- 0.010, substantially above a logistic prompt-length baseline of 0.657 +/- 0.014, and none of the nonlinear models tested improves upon it. A companion question about whether self-repair leaves a geometric signature in the model's hidden states could not be answered, because successful repairs following a failed first attempt are too rare in this setting to support the analysis. The contribution is both empirical and methodological, providing evidence that pre-generation hidden states contain a robust signal of eventual code correctness, together with a confound-control diagnostic that quantifies how much of that signal survives adjustment for prompt length.

Carlo Di Cicco

Independent researcher.

Code Language Models LLM Code Generation Hidden-State Probing

Jun 11, 2026cs.LG

Detecting Functional Memorization in Code Language Models

Large language models (LLMs) are increasingly used to generate code at scale. Meanwhile, prior work has investigated whether training data may be recoverable from model outputs, by auditing the textual overlap between training examples and model generations. Code, however, can be functionally equivalent while textually dissimilar. In this work, we study functional memorization: extraction of functional logic beyond what verbatim metrics detect. We construct a counterfactual setup for Olmo-3-32B, comparing a midtrained model (exposed to target code) against a pretrained reference (not exposed). We prompt both models with Python function signatures and measure both textual and functional similarity (i.e., LLM-as-a-judge, execution-based). Our results show clear evidence of functional memorization, highlighting the need for auditing metrics that go beyond textual overlap.

Matthieu Meeus, Anil Ramakrishna, Matthew Grange +2

Meta · Imperial College London

Memorization in Language Models LLM Auditing Memorization in Generative Models Code Language Models

Jun 10, 2026cs.LG

CODEBLOCK: Learning to Supervise Code at the Right Granularity

Supervised fine-tuning of code LLMs typically applies uniform cross-entropy loss to all response tokens, implicitly assuming that every token provides equally useful learning signals. Recent token-level selection methods challenge this assumption in natural-language SFT by supervising only high-value tokens. However, such pointwise selection can fragment the syntactic structures and program dependencies of code, leaving supervision scattered across incomplete code units. In experiments on 30K code instruction-response pairs, under the same 10% token budget, supervising complete coding blocks improves average performance by 11.7 points over prior approaches that supervise isolated tokens. Motivated by this observation, we propose CodeBlock, a structure-aware sparse supervision framework that uses complete, parser-aligned coding items as the basic units of supervision. CodeBlock constructs coding items from high-quality code instruction data, estimates their supervision utility using GCE, which is more robust to low-probability tokens, and further adjusts their supervision priority using data-flow reach and bridge signals. During training, the full response is retained as context, while loss is applied only to the selected supervision units. Experiments show that partial supervision over only about 6.9% of response tokens consistently outperforms full-token SFT across all five model settings, suggesting that effective code SFT depends not only on identifying high-value supervision, but also on allocating it at the appropriate structural granularity.

Zhijie Deng, Ling Li, Jinlong Pang +5

Hong Kong University of Science and Technology (Guangzhou) · UC Santa Cruz · Ant Group +1

Supervised Fine-Tuning Code Language Models LLM Code Generation

Jun 8, 2026cs.CL

CodeAlchemy: Synthetic Code Rewriting at Scale

Pre-training on raw code teaches syntax but provides sparse signal for diverse real-world task formats. While synthetic data has proven transformative for language models, code remains largely unexplored beyond limited quality improvements. We present CodeAlchemy, a synthetic data generation framework that transforms publicly sourced code into semantically-rich training data through 5 strategies: CodeEnhance (quality-aware rewriting), CodeQA (template-based problems), CodeDev (developer tasks), CodeDialogue (multi-turn conversations), and CodeTrace (execution traces). We process 3 corpora across 15 languages to generate 500B+ tokens of synthetic data plus 350B reasoning tokens, orders of magnitude more than prior efforts. CodeTrace instruments and executes 1.3M+ files across 14 languages and 5K libraries, capturing control flow, state tracking, and library knowledge. We introduce DevEval (developer tasks) and TraceEval (execution prediction) benchmarks; frontier models like Claude Sonnet 4.5 achieve only 5.6% exact match on TraceEval, revealing critical gaps in semantic understanding. Our 3B models achieve 83.5% on HumanEval, 63.2% on MBPP, 8.09% win rate on DevEval, and 15.36 ROUGE-2 on TraceEval, outperforming frontier models 10x the size including 27B Gemma-3 and 32B Granite-4.0.

Ankit Gupta, Aditya Prasad, Rameswar Panda

MIT-IBM Watson AI Lab, IBM Research

Synthetic Data Pretraining Synthetic Data Generation Code Language Models Code Generation Benchmarks

Jun 8, 2026cs.CL

Code Is More Than Text: Uncertainty Estimation for Code Generation

Large language models (LLMs) are increasingly deployed as code generators, where silently wrong programs pose real safety and reliability risks. Reliable uncertainty estimation (UE) is essential for selective prediction, human-in-the-loop review, and downstream agentic decisions. Yet most existing code UE methods are inherited from natural language (NL) generation and ignore properties that make code distinct. We argue that code differs from NL in three ways: a single wrong token can break an entire program (token fragility); algorithmic intent and concrete implementation can disagree independently (intent-code gap); and programs can be executed (executability). We instantiate these properties as three orthogonal uncertainty axes: lexical (Top-K token entropy), algorithmic (pseudo-code consistency), and functional (behavioral consistency). Across five code LLMs, our three-axis ensemble improves average AUROC from 0.696 for the strongest NL-derived baseline to 0.776 (+8.1 points). Notably, on Qwen3-14B, our single-pass Top-K token entropy matches the strongest multi-pass baseline while being over 3x cheaper; across models, it remains a competitive low-cost signal. These results suggest that code UE deserves code-specific design rather than direct NL ports.

Yuling Shi, Caiqi Zhang, Yuexian Li +4

Shanghai Jiao Tong University · University of Cambridge

Confidence Estimation in Language Models LLM Reliability LLM Uncertainty Estimation Code Language Models LLM Code Generation

Jun 7, 2026cs.AI

Beyond Pass Rate: A Multilingual, Execution-Grounded Evaluation of Open Code LLMs

Code generation models are typically compared using compact execution benchmarks and aggregate pass rates, but such summaries obscure how performance varies across programming languages, problem families, and failure modes. We present a large-scale, execution-grounded evaluation of 9 openly accessible LLMs specialized for coding on 2,707 free LeetCode problems across 12 programming languages. Our corpus contains 325,343 problem-model-language jobs, each linked to prompt metadata, extracted code, LeetCode execution outcomes, and static-analysis signals. The results show that current open models remain far from the human acceptance reference: the best model, Yi-Coder-9B-Chat, reaches 23.64% mean correctness, compared with a 57.2% human acceptance baseline. Rankings are also slice-dependent: Qwen2.5-Coder-14B-Instruct is strongest on hard problems and distinct-problem coverage, while Gemma-2-27B-IT achieves the highest all-language lint pass rate. Failure analysis shows that compile errors account for 63.25% of non-accepted best submissions, indicating that many failures occur before semantic correctness can be tested. Static quality further diverges from functional correctness. Together, these findings show that multilingual, artifact-preserving evaluation reveals tradeoffs hidden by single-language or single-metric leaderboards.

Sayed Erfan Arefin

Department of Computer Science University of Dayton Dayton, OH 45469, USA

Code Generation Evaluation Code Language Models Code Generation Benchmarks Multilingual Code Generation

Jun 7, 2026cs.SE

Lost in the Flow with Code Talkers: Unveiling the Instruction-Tuning Tax of Large Language Models in Code Tasks

AI coding assistants have significantly improved developer productivity by automatically suggesting code that aligns with user intent, and many of these tools are now integrated directly into Integrated Development Environments (IDEs). Developers interact with code in two distinct cognitive modes: Flow and Command. While developers require tools that directly complete or infill code in unfinished programs during Flow mode, they also need tools that can comprehend intentions expressed as natural-language instructions and convert them into executable code in Command mode. Although instruction-tuned Large Language Models (LLMs) dominate many application scenarios due to their abilities to infer and fulfill developers' intents, it remains unclear whether the same paradigm is equally suitable for different code-related tasks. Therefore, it is necessary to understand how instruction tuning affects the feasibility of CodeLLMs as coding assistants. To fill this gap, we conduct the first empirical study that uncovers a key trade-off caused by instruction tuning across programming modes, which we term the Instruction-Tuning Tax. Our results show that instruction tuning is not a free lunch: although instruction-tuned models are more capable of following instructions and leveraging structured guidance, these gains often come at the cost of weaker infilling performance. We further extend our study through both qualitative and quantitative analyses, including manual failure categorization, behavioral metrics that capture generation fidelity, and intermediate-checkpoint evaluation throughout the tuning process. Summarizing our results into seven findings and four implications, our study offers a new perspective on the development of AI-powered coding tools and highlights the need to carefully balance instruction-following ability with effective code generation assistance.

Shi Ying Chang, Chiok Yew Ho, Yichen Li +1

The Chinese University of Hong Kong, China

Instruction Tuning Code Generation Evaluation Code Language Models LLM Code Generation

Jun 4, 2026cs.SE

Code2LoRA: Hypernetwork-Generated Adapters for Code Language Models under Software Evolution

Code language models need repository-level context to resolve imports, APIs, and project conventions. Existing methods inject this knowledge as long inputs (retrieved through RAG or dependency analysis) or through per-repository fine-tuning and LoRA -- costly at repository scale and brittle to evolving codebases. We introduce Code2LoRA, a hypernetwork framework that generates repository-specific LoRA adapters, effectively injecting repository knowledge with zero inference-time token overhead. Code2LoRA supports two usage scenarios: Code2LoRA-Static converts a single repository snapshot into an adapter, suitable for comprehension of stable codebases; while Code2LoRA-Evo maintains an adapter backed by a GRU hidden state updated per code diff, suitable for active development of evolving codebases. To evaluate Code2LoRA against parameter-efficient fine-tuning baselines, we build RepoPeftBench, a benchmark of 604 Python repositories with two tracks: a static track with 40K training and 12K test assertion-completion tasks, and an evolution track with 215K commit-derived training and 87K commit-derived test tasks. On the static track, Code2LoRA-Static achieves 63.8% cross-repo and 66.2% in-repo exact match, matching the per-repository LoRA upper bound; on the evolution track, Code2LoRA-Evo achieves 60.3% cross-repo exact match (+5.2 pp over a single shared LoRA). Code2LoRA's code can be found at https://anonymous.4open.science/r/code2lora-6857; the model checkpoints and RepoPeftBench datasets can be found at https://huggingface.co/code2lora.

Liliana Hotsko, Yinxi Li, Yuntian Deng +1

University of Waterloo

Software Engineering LLM Fine-Tuning Code Language Models Parameter-Efficient Fine-Tuning

Jun 2, 2026cs.AI

NovelAPIBench: Diagnosing How A Code LLM Learns to Use Novel APIs

Rapidly evolving software libraries require coding language models to use unfamiliar APIs. Yet failed solutions alone tell us little about what API knowledge models need or how to provide it. We present NovelAPIBench, an automated diagnostic framework that builds benchmarks around the API knowledge gaps of a target model. It identifies novel APIs, generates executable coding tasks with separately controllable knowledge components, and classifies failures into six categories. The framework can be rerun as models and libraries evolve. We study 21 libraries across five domains and six backbone models, with our primary benchmark comprising 1,670 tasks covering 856 APIs. Our experiments show that when supplied directly, usage examples provide the strongest standalone guidance for novel API acquisition, and combining them with signature descriptions approaches full-knowledge performance. Retrieving this knowledge from a pool of APIs yields lower accuracy than supplying only the target API's knowledge, even when retrieval finds the target. Fine-tuning on other APIs improves performance on unseen APIs mainly when external knowledge is supplied, helping models both select the target API and integrate it into the surrounding code. These findings highlight complementary roles for external API knowledge and the ability to apply it. By separating knowledge content, delivery, and use, NovelAPIBench helps diagnose what limits novel API use and guides adaptation to evolving libraries. Code and data are available at https://github.com/MAPS-research/NovelAPIBench.

Jinnuo Liu, Yue Peng, Jinhan Niu +1

NYU Shanghai

Knowledge Augmentation for Language Models LLM Fine-Tuning Code Language Models LLM Code Generation

May 27, 2026cs.CR

Code as a Weapon: A Consensus-Labeled Prompt Bank for Measuring Coding-Model Compliance with Malicious-Code Requests

A general-purpose language model that answers a harmful question returns text; a coding model that complies with a malicious request can return a working weapon: a keylogger, ransomware, an exploit that runs as written. This asymmetry in the severity of a single act of compliance implies coding-specialized models should clear a higher refusal bar than general-purpose chat models, not a lower one, yet the field cannot tell whether they do. Refusal benchmarks for malicious code are fragmented: they mix requests for executable software with requests for harmful security knowledge and report refusal rates over non-comparable corpora. This paper's central result is that the CODE-versus-KNOWLEDGE classification axis established in a prior four-corpus release remains stable under a substantially expanded corpus pool and an independently refreshed judge panel, evidence that it measures a real construct rather than an artifact of the prompts or judges. Eight corpora spanning diverse elicitation paradigms (direct, jailbreak-decorated, indirect, and agent/interpreter: ASTRA, CySecBench, AdvBench/harmful_behaviors, JailbreakBench, MalwareBench, RedCode, RMCBench, Scam2Prompt) are classified under a five-judge consensus protocol (6,675 prompts x 5 judges = 33,375 calls), reaching Fleiss' kappa = 0.767 [95% CI 0.755, 0.777] ("substantial"). Critically, the panel shares no judge with the prior release (five paid commercial APIs replaced by five open-weight models from five vendors), yet the two panels agree on 94.45% of the 3,133 shared prompts and reach Cohen's kappa = 0.952 [0.942, 0.963] on the 3,031-prompt binary overlap: the axis survives near-total panel replacement. The released bank comprises 4,748 consensus-CODE and 1,923 consensus-KNOWLEDGE prompts, a reliability-quantified benchmark whose central classification axis is shown stable across corpus expansion and judge-panel replacement.

Richard J. Young, Gregory D. Moody

University of Nevada Las Vegas · Department of Information Systems · Las Vegas, NV, USA

Classification LLM Safety Evaluation Benchmark Validity Refusal Behavior in Language Models Code Language Models

May 26, 2026cs.CR

Poison with Style: A Practical Poisoning Attack on Code Large Language Models

Code Large Language Models (CLLMs) serve as the core of modern code agents, enabling developers to automate complex software development tasks. In this paper, we present Poison-with-Style (PwS), a practical and stealthy model poisoning attack targeting CLLMs. Unlike prior attacks that assume an active adversary capable of directly embedding explicit triggers (e.g., specific words) into developers' prompts during inference, PwS leverages developers' code styles as covert triggers implicitly embedded within their prompts. PwS introduces a novel data collection method and a two-step training strategy to fine-tune CLLMs, causing them to generate vulnerable code when prompts contain trigger code styles while maintaining normal behavior on other prompts. Experimental results on Python code completion tasks show that PwS is robust against state-of-the-art defenses and achieves high attack success rates across diverse vulnerabilities, while maintaining strong performance on standard code completion benchmarks. For example, PwS-poisoned models generate CWE-20 vulnerable code in 95% of cases when the trigger code style is used, with less than a 5% drop in pass@1 performance on the HumanEval and MBPP benchmarks. Our implementation and dataset are here: https://github.com/khangtran2020/pws.

Khang Tran, Yazan Boshmaf, Issa Khalil +3

Department of Data Science, New Jersey Institute of Technology, New Jersey, U.S.A. · Qatar Computing Research Institute, HBKU, Doha, Qatar · Mohamed Bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates

LLM Backdoor Attacks Data Poisoning Attacks Code Language Models LLM Code Generation

May 25, 2026cs.LG

Merge-Bench: Resolve Merge Conflicts with Large Language Models

This paper applies machine learning to the difficult and important task of version control merging. (1) We constructed a dataset, Merge-Bench, of 7938 real-world merge conflict hunks from 1439 GitHub repositories. The ground truth is the merge resolution that developers committed to the repository. Our dataset construction methodology is scalable to arbitrary amounts of data since no manual labeling is required. (2) We trained a model, LLMergeJ, to resolve merge conflicts in Java programs. Our approach uses Group Relative Policy Optimization (GRPO), an online reinforcement learning method, to train a Large Language Model (LLM). (3) We performed two evaluations of the performance of LLMs on resolving merge conflicts. On Java programs, LLMergeJ with 14B parameters outperforms 3 commercial LLMs, trailing only Gemini 2.5 Pro. Across 11 programming languages, commercial LLM performance is largely stable from language to language. The best models correctly resolve less than 60% of merge conflicts.

Benedikt Schesch, Michael D. Ernst

Amazon · University of Washington, USA

Software Engineering Code Language Models LLM Training

May 23, 2026cs.CL

CSP-Atlas: Concept-Specific Neural Circuits in a Sparse Python Transformer

A sparse 8-layer code transformer develops dedicated neural circuitry for every Python construct tested, and that circuitry is organised by a clean computational principle rather than by semantic category. We extract neural circuits for 106 concepts (43 AST node types, 63 builtin objects) by marginalising across 63,800 controlled prompts, and decompose each circuit into concept-specific and token-driven components using contrastive checker prompts that present a keyword token without its associated syntactic structure. Three findings emerge. First, all 106 concepts produce non-empty universal circuits at every one of nine parameter settings, and the ranking of concept-specificity across constructs is stable across the sweep - survival is not an artifact of a permissive threshold. Second, AST circuits contain a genuine concept component distinct from token activation: concept-only neurons constitute up to 62.5% of the loudest-firing neurons at mid-to-late layers, while builtin circuits are almost entirely token-driven. Third, six computationally atomic constructs - Import, ImportFrom, Break, Continue, Pass, Assert - cluster together despite being semantically unrelated, sharing only the property of being single-statement constructs requiring no nested body; this atomicity super-cluster, together with a four-tier hierarchy organised by token ambiguity and structural distinctiveness, shows that the model's internal organisation tracks computational structure rather than meaning. The methodology, full decomposition data, and analysis code are released.

Piotr Wilam

University College London

Transformer Interpretability Circuit Discovery Mechanistic Interpretability Circuits in Language Models Code Language Models

May 22, 2026cs.SE

TRACER: A Semantic-Aware Framework for Fine-Grained Contamination Detection in Code LLMs

Data contamination is a known threat to the reliability of model evaluation. However, it remains underexplored in code large language models (LLMs), where contamination often goes beyond exact duplication. We present TRACER, a semantic-aware framework for fine-grained code contamination detection. TRACER models contamination using three levels of semantic overlap - Functionally Identical, Nearly Identical, and Shared Logic - and detects them through a coarse-to-fine pipeline. We also introduce the first benchmark for fine-grained code contamination detection, spanning three widely used benchmarks and three representative post-training datasets. TRACER achieves strong and consistent performance across multiple LLM backbones, with GPT-5 reaching an F1 score of 0.91 in fine-grained detection. In the binary setting, TRACER attains an F1 of 0.92, outperforming existing methods by 42%-217%. We further conduct ablation studies and error analysis to assess the contributions of individual components in TRACER.

Yifeng Di, Xuliang Huang, Tianyi Zhang

Purdue University West Lafayette, IN · University of Illinois Urbana-Champaign Champaign, IL

LLM Auditing Data Contamination in Language Models Code Language Models

May 21, 2026cs.AI

Echo: Learning from Experience Data via User-Driven Refinement

Static "human data" faces inherent limitations: it is expensive to scale and bounded by the knowledge of its creators. Continuous learning from "experience data" - interactions between agents and their environments - promises to transcend these barriers. Today, the widespread deployment of AI agents grants us low-cost access to massive streams of such real-world experience. However, raw interaction logs are inherently noisy, filled with trial-and-error and low information density, rendering them inefficient for direct model training. We introduce Echo, a generalized framework designed to operationalize the transition from raw experience to learnable knowledge, effectively "echoing" environmental feedback back into the training loop for model optimization. In today's agent ecosystem, user refinement serves as a primary source of such feedback: driven by responsibility for the outcome, users rigorously transform flawed agent proposals into verified solutions. These user-driven refinement sequences inherently distill agents' crude attempts into high-quality training signals. Echo systematically harvests these signals to continuously align the agent with real-world needs. Large-scale validation in a production code completion environment confirms that Echo effectively harnesses this pipeline, breaking the static performance ceiling by increasing the acceptance rate from 25.7% to 35.7%.

Hande Dong, Xiaoyun Liang, Jiarui Yu +15

Tencent, CodeBuddy

Continual Learning for LLM Agents LLM Agent Self-Improvement Human-in-the-Loop AI Code Language Models RL from Human Feedback

May 16, 2026cs.CR

The Range Shrinks, the Threat Remains: Re-evaluating LLM Package Hallucinations on the 2026 Frontier-Model Cohort

Spracklen et al. (USENIX Security '25) showed that code-generating large language models hallucinate package names that do not exist on PyPI or npm at rates ranging from 5.2% on commercial models to 21.7% on open-source models, creating an attack surface for slopsquatting -- the registration of malicious packages under hallucinated names. We replicate their methodology on five frontier code-capable LLMs released between October 2025 and March 2026: Claude Sonnet 4.6, Claude Haiku 4.5, GPT-5.4-mini, Gemini 2.5 Pro, and DeepSeek V3.2. Across 199,845 paired Python and JavaScript prompts validated against PyPI and npm master lists, we measure overall hallucination rates between 4.62% (Claude Haiku 4.5) and 6.10% (GPT-5.4-mini) -- an order-of-magnitude compression of the inter-model spread observed by Spracklen, but not a retirement of the threat. Beyond replication, we identify a set of 127 package names (109 on PyPI, 18 on npm) that all five evaluated models invent identically; following coordinated disclosure with PyPI Security and Socket.dev, 53 of these (41 on PyPI, 12 on npm) remain registrable by an attacker after each registry's existing defenses, constituting a model-agnostic supply-chain attack surface that no single-model study can reveal. We further document a Python-over-JavaScript hallucination asymmetry that inverts Spracklen's 2024 finding, identify a Haiku-below-Sonnet inversion within the Anthropic family, and observe a Jaccard-similarity peak between DeepSeek V3.2 and GPT-5.4-mini (J = 0.343) suggestive of shared training-data origins.

Aleksandr Churilov

Independent Researcher

LLM Evaluation Hallucination in Language Models Software Supply Chain Security Code Language Models LLM Code Generation

May 15, 2026cs.CL

Syntax Without Semantics: Teaching Large Language Models to Code in an Unseen Language

Large language models (LLMs) achieve high pass rates on code generation benchmarks, yet whether they can transfer this ability to languages absent from pretraining remains poorly understood. We introduce PyLang, a minimal imperative language absent from all pretraining corpora, and evaluate frontier models zero-shot and fine-tuned Qwen3 (4B, 8B, 32B) on 352 problems. We find that fine-tuning quickly teaches syntax but fails to transfer semantic competence: Python outperforms PyLang by up to 19% across all configurations, and no intervention (multi-task learning, preference tuning, code infilling, or latent-space objectives) closes the gap. An LLM judge reveals that frontier models select an identical algorithm to Python 80% of the time, yet cannot translate it into a working PyLang implementation., and CKA analysis confirms that fine-tuned models converge to nearly identical internal representations across languages (CKA > 0.97) while diverging at the output stage. We term this the implementation fidelity gap: models possess language-agnostic algorithmic understanding but cannot express it in an unfamiliar language. Our findings highlight the need for training methods that decouple reasoning from language-specific realization.

Vinayshekhar Bannihatti Kumar, Disha Makhija, Manoj Ghuhan Arivazhagan +1

AWS AI Labs

LLM Fine-Tuning Code Language Models LLM Code Generation

May 14, 2026cs.SE

Correctness-Aware Repository Filtering Under Maximum Effective Context Window Constraints

Context window efficiency is a practical constraint in large language model (LLM)-based developer tools. Paulsen [12] shows that all tested models degrade in accuracy well before their advertised context limits the Maximum Effective Context Window (MECW) which makes context construction a quality problem, not just a cost one. Modern software repositories routinely contain large non-code artifacts compiled datasets, binary model weights, minified JavaScript bundles, and gigabyte-scale log files that overflow the context window and push out task-relevant source code. We present a correctness-aware context hygiene framework: a pre-execution, size-based heuristic filter that intercepts repository scans before tokenization, using only OS-level stat() metadata with sub-millisecond overhead. Semantic retrieval approaches such as RepoCoder, GraphRAG, and AST-based chunking require index construction and query-time inference before any filtering decision is reached. Our framework, by contrast, requires no indexing and operates at <0.01 ms per file decision. Across 10 real open-source repositories (22,046 files, 5 languages), the proposed SizeFilter at θ=1 MB achieves 79.6% (\pm13.2%) mean token reduction at 0.30 ms overhead: the HybridFilter achieves 89.3% (\pm9.0%) the lowest variance of any filter evaluated. A token-density study across 2,688 files confirms a strong linear correlation (Pearson r=0.997, k=0.250 tokens/byte). A limited-scope evaluation (18 tasks, CodeLlama-7B-Instruct) yields 72% file-level accuracy under filtering versus 25% at baseline; hallucination frequency declines from 61% to 17%. All code and data are released for reproducibility.

Shweta Mishra

Independent Researcher

LLM Context Management Code Language Models

May 13, 2026cs.SE

Code-Centric Detection of Vulnerability-Fixing Commits: A Unified Benchmark and Empirical Study

Automated detection of vulnerability-fixing commits (VFCs) is critical for timely security patch deployment, as advisory databases lag patch releases by a median of 25 days and many fixes never receive advisories. We present a comprehensive evaluation of code language model based VFC detection through a unified framework consolidating over 20 fragmented datasets spanning more than 180000 commits. Across over 180 experiments with fine-tuned models from 125 M to 14 B parameters, we find no evidence that models acquire transferable security-relevant code understanding from code changes alone. When commit messages are available, they dominate model attention, and when removed, an attribution analysis shows that enriching diffs with additional intra-procedural semantic context does not shift model attention toward the code changes. Group-stratified evaluation exposes approximately 17% performance drops compared to random splits, while temporal splits on aggregated datasets prove unreliable due to compositional shift in the underlying project distributions. At a false positive rate of 0.5% all fine-tuned code-only models miss over 93% of vulnerabilities. Larger and more diverse training data or generative approaches show preliminary improvements but do not resolve the underlying limitations. To support future research on code-centric VFC detection, we release our unified framework and evaluation suite.

Nils Loose, Joseph Bienhüls, Kristoffer Hempel +2

University of Lübeck Institute for IT Security Lübeck, Germany

Software Vulnerability Detection Software Engineering Benchmarks Code Language Models

May 12, 2026cs.SE

SemChunk-C: Semantic Segmentation for C Code

Semantic segmentation of code written in a C-family language remains a challenging problem, due to the language's complex syntax, macro expansion, and irregular structural patterns. Existing chunking methods, such as fixed-sized windows, heuristic splitting, and syntax-based tools, often fail to capture meaningful functional units, limiting the efficacy of retrieval and other downstream LLM driven tasks. In this paper, we address the problem of chunking in C-related languages. First, we define a set of code chunk categories. Second, we train an LLM-based classifier to a) identify chunk boundaries, and b) assign each chunk a descriptive functional attribute (a category), which can be useful for downstream tasks. By leveraging the LLM's ability to capture semantic context within the code, we assume flexible chunk boundaries, allowing to adapt to the specific structure and context of each instance. Third, we introduce SemChunk-C, a family of lightweight language models for semantic chunking of C-related files (.c, .cpp, .h, .cs, etc.). These models are based on the first four Ettin encoders [1] with 17M, 32M, 68M, and 150M parameters. Despite their relatively small size, they are capable of identifying cohesive code units, such as data structures, interface blocks, and other components. Furthermore, we demonstrate the robustness of our approach on real-world code, including challenging constructs such as nested definitions and macros. We test our approach on various datasets, and show that it achieves high boundary accuracy and semantic coherence, matching or outperforming chunkers that are based on much larger code-oriented LLMs. We also validate the improved performance of the downstream tasks on a few curated benchmarks.

Boris Nazarov, Darya Frolova, Shaked Leibzirer +1

Small Language Models Code Language Models