The results recast package hallucination as both a measurement problem and a decoding-time control problem, and they demonstrate that the choice of defense must be matched to the threat model and recommendation utility.
Abstract
LLMs are increasingly used for code generation, yet they frequently hallucinate non-existent software packages, creating exploitable entry points into the software supply chain. We make four contributions to this problem. First, we show that prior evaluation methodologies systematically inflate hallucination rates by misclassifying standard-library modules as hallucinations in some languages. For Python, the overestimation reaches 9.4 percentage points. Second, we evaluate seven inference-time defenses for mitigating package hallucinations, including five guided decoding strategies (Greedy, Contrastive, DoLa, Nudging, and Active Layer-Contrastive Decoding), an iterative self-refinement approach (Self-Refine), and a Retrieval-Augmented Generation (RAG)-based defense.. Across eight models spanning five families and four programming languages (Python, JavaScript, Ruby, Rust), RAG reduces the package hallucination rate (PHR) in 18 of 32 model--language configurations. Third, we introduce Package Utility (PU) to assess whether defenses preserve valid and task-relevant recommendations. Among strategies evaluated, Greedy decoding provides the strongest average mitigation--utility trade-off. Fourth, we stress-test all strategies under adversarial prompts seeded with fabricated package names and find that PHR surges by up to 45 percentage points relative to standard prompts, with Ruby consistently the most vulnerable language (80.9--95.2\%). Under adversarial conditions, RAG and Self-Refine outperform all decoding-only strategies, indicating that robust defense requires either external grounding or iterative self-verification when prompts are actively hostile. Our results recast package hallucination as both a measurement problem and a decoding-time control problem, and they demonstrate that the choice of defense must be matched to the threat model and recommendation utility.
EAHC is proposed, an Execution-Anchored Hallucination Calibration reranking framework that anchors reasoning judgments to execution behavior so that execution-equivalent candidates receive consistent scores, which implements a dual-channel architecture.
Guang Yang, Xing Hu, Xiang Chen et al.· 0 citations
Large language models (LLMs) frequently produce source code that seems correct and well-formed, yet includes hallucinated elements that cause downstream test failures. In this study, we benchmark state-of-the-art uncertainty quantification methods and existing base-lines for the task of hallucination detection in source code and introduce a diff-based pipeline to construct a code dataset annotated with line-level hallucinations. Building on this, we train a lightweight Transformer-based detector that uses LLM internal representations to identify hallucinations, substantially outperforming existing methods across several code generation domains. The detector also shows particular promise for enabling self-correction in LLM-based coding agents. We release the first publicly available dataset of line-level code hallucinations, along with the corresponding source code and trained hallucination detectors https://github.com/ datapaf/CodeHallucinationDetection
G. Andriushchenko, Roman Garaev, L. Rvanova et al.· Annual Meeting of the Associ...· 0 citations
Unit testing plays a critical role in ensuring software quality and reliability in large-scale industrial environments. While Large Language Models (LLMs) offer promising automated test generation capabilities, their practical deployment faces significant challenges due to hallucination problems. In this paper, we analyze compilation failures from LLM-generated unit tests in Ant Group's production systems and identify two fundamental types of hallucinations: extrinsic hallucinations caused by insufficient contextual information and intrinsic hallucinations stemming from model limitations even with adequate context. To address these issues, we propose DEHALL, an automated end-to-end unit test generation tool that systematically mitigates both types of hallucinations through comprehensive context construction and targeted static analysis-based repair. Our approach builds a heterogeneous graph to capture relevant context and employs specialized repair mechanisms for import, field, and method issues. Evaluation on Ant Group's internal datasets reveals that DEHALL achieves 71.56% line coverage and 67.18% branch coverage, significantly outperforming vanilla LLM approaches. In the public benchmarks, it also shows better performance on coverage and better defect detection capability than previous state-of-the-art approaches. DEHALL has been successfully deployed across multiple business domains at Ant Group, achieving an 81% developer adoption rate with positive user feedback on productivity improvements.
HallDetect, a lightweight, reference-free, and black-box framework for hallucination detection, is presented, a lightweight, reference-free, and black-box framework for hallucination detection that is evaluated not only on summarization but across a broader range of source-grounded generation settings.
Achir Oukelmoun, N. Semmar, Gäel de Chalendar· 0 citations
In code generation tasks, large language models (LLMs) frequently recommend nonexistent software packages, referred to as package hallucinations. This may pose a security risk, as attackers could register these fake packages and exploit them in supply chain attacks. To investigate the real-world feasibility, this paper conducts a systematic empirical study of this potential threat. We extracted 107 hallucinated packages from eight prominent LLMs and uploaded them to the npm registry over six months. The results demonstrate a tangible risk: 35.2% of package installations are attributed to developers, with most initial interactions occurring within a two-week window post-registration. A name similarity analysis further reveals that 25.2% of hallucinated names fall within an absolute edit distance of \(\leq\) 1 from existing npm packages, posing a direct typosquatting risk. In our evaluated setting, tool-augmented generation, specifically using web search, fails to mitigate this threat. Instead, it significantly amplifies hallucination persistence from 11.2% to 17.3% through context contamination. In the npm registry, the entire attack campaign costs only an estimated $496, making it far more accessible than traditional supply chain attacks yet capable of enabling poisoning effects. To mitigate this emergent threat, we design an autonomous defense agent: NymGuard. It provides a proactive defense by automating the real-time detection of hallucinated packages, their preemptive registration to block malicious use and the maintenance of a public warning registry. The full implementation has been open-sourced and is available on GitHub (https://github.com/xiaoye798/Package-Hallucination-Research).
Zhe Huang, Chaoge Liu, Zhi Wang et al.· ACM Transactions on Software...· 0 citations
This paper describes a system for the SemEval-2026 Task 12 on Abductive Event Reasoning (AER). We systematically address the pervasive “over-selection” hallucination pathology in Instruction-tuned Large Language Models (LLMs), where generative models erroneously align distractor options with semantic priors rather than retrieved evidence. The proposed architecture utilizes a 32-billion-parameter foundational model within a Retrieval-Augmented Generation (RAG) pipeline. To combat the hallucination bottleneck and adapt to the strict penalty for incorrect predictions, we propose a Precision-Oriented Decoding (POD) strategy, which tightly couples low-temperature Zero-shot Chain-of-Thought (CoT) sampling with scaled marginalization (majority voting). De-ployed efficiently on consumer-grade hardware via Low-Rank Adaptation (LoRA) fine-tuning, our system achieved a highly competitive average score of 0.802 on the official test set. Through a rigorous three-stage empirical evolution and analysis of the diagnostic failure of an asymmetric logical ensemble, the experimental results demonstrate that epistemic noise suppression is strictly superior to heuristic recall compensation in abductive reasoning tasks.
Chiyao Zhou, Zebing Wang, Kexin Deng et al.· SemEval@ACL· 1 citation
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduAug 17, 2026
A USAF cadet and a Lincoln Laboratory researcher found AI chatbots can help nontechnical service members produce viable software applications for their unique problems.
