arXiv subjects

Micheline Bénédicte Moumoula

Publications and source records attributed to Micheline Bénédicte Moumoula.

5 recordsLinked to original sources

arXiv2026

Programming Language Confusion: When Code LLMs Can't Keep their Languages Straight

Large Language Models (LLMs) have achieved state-of-the-art performance across software engineering tasks, from code generation to translation. However, we identify and systematically evaluate a critical failure mode: Programming Language Confusion (PLC) -- the generation of code in unintended languages despite explicit instructions. Through evaluation of 10 popular LLMs across six multilingual datasets (LiveCodeBench, BabelCode variants, HumanEval-XL, and McEval), we demonstrate that PLC is pervasive, with some specialized models exhibiting the highest confusion rates. Our analysis reveals that PLC is not random noise but reflects systematic patterns: models consistently generate syntactically valid code even when it deviates from language specifications. This behavior produces distinct language migration patterns, most notably a strong default to Python and systematic shifts between syntactically similar language pairs (e.g., C#/Java). These migrations reflect statistical preferences learned from training data rather than goal-directed reasoning. We demonstrate that explicit language keywords provide the most effective mitigation, while natural language instructions have limited influence on model behavior. Furthermore, model quantization -- though essential for practical deployment -- significantly amplifies PLC and degrades syntactic stability in complex tasks. Our findings underscore that language fidelity should be treated as a core evaluation dimension for code LLMs. We advocate for standardized benchmarks and prompt formats with explicit language constraints to enable more reliable assessment and foster the development of robust, multilingual code generation systems.

cs.SE↗

arXiv2025

Beyond Language Barriers: Multi-Agent Coordination for Multi-Language Code Generation

Producing high-quality code across multiple programming languages is increasingly important as today's software systems are built on heterogeneous stacks. Large language models (LLMs) have advanced the state of automated programming, yet their proficiency varies sharply between languages, especially those with limited training data such as Rust, Perl, OCaml, and Erlang. Many current solutions including language-specific fine-tuning, multi-agent orchestration, transfer learning, and intermediate-representation pipelines still approach each target language in isolation, missing opportunities to share knowledge or exploit recurring cross-language patterns. XL-CoGen tackles this challenge with a coordinated multi-agent architecture that integrates intermediate representation, code generation, translation, and automated repair. Its distinguishing feature is a data-driven mechanism for selecting bridging languages: empirically derived transfer matrices identify the best intermediate languages based on demonstrated translation success rather than raw generation accuracy. The system performs early output validation, iteratively corrects errors, and reuses intermediate artifacts as contextual scaffolds for subsequent translations. Extensive experiments show that XL-CoGen yields notable improvements with 13 percentage-point gains over the strongest fine-tuned baseline and as much as 30 percentage points over existing single-language multi-agent methods. Ablation studies further demonstrate that compatibility-guided bridging significantly outperforms LLM-based heuristics, confirming the value of cumulative cross-language knowledge transfer.

cs.SE↗

arXiv2025

How Small Transformation Expose the Weakness of Semantic Similarity Measures

This research examines how well different methods measure semantic similarity, which is important for various software engineering applications such as code search, API recommendations, automated code reviews, and refactoring tools. While large language models are increasingly used for these similarity assessments, questions remain about whether they truly understand semantic relationships or merely recognize surface patterns. The study tested 18 different similarity measurement approaches, including word-based methods, embedding techniques, LLM-based systems, and structure-aware algorithms. The researchers created a systematic testing framework that applies controlled changes to text and code to evaluate how well each method handles different types of semantic relationships. The results revealed significant issues with commonly used metrics. Some embedding-based methods incorrectly identified semantic opposites as similar up to 99.9 percent of the time, while certain transformer-based approaches occasionally rated opposite meanings as more similar than synonymous ones. The study found that embedding methods' poor performance often stemmed from how they calculate distances; switching from Euclidean distance to cosine similarity improved results by 24 to 66 percent. LLM-based approaches performed better at distinguishing semantic differences, producing low similarity scores (0.00 to 0.29) for genuinely different meanings, compared to embedding methods that incorrectly assigned high scores (0.82 to 0.99) to dissimilar content.

cs.CL↗

arXiv2025

Using Contrastive Learning to Improve Two-Way Reasoning in Large Language Models: The Obfuscation Task as a Case Study

This research addresses a fundamental question in AI: whether large language models truly understand concepts or simply recognize patterns. The authors propose bidirectional reasoning,the ability to apply transformations in both directions without being explicitly trained on the reverse direction, as a test for genuine understanding. They argue that true comprehension should naturally allow reversibility. For example, a model that can change a variable name like userIndex to i should also be able to infer that i represents a user index without reverse training. The researchers tested current language models and discovered what they term cognitive specialization: when models are fine-tuned on forward tasks, their performance on those tasks improves, but their ability to reason bidirectionally becomes significantly worse. To address this issue, they developed Contrastive Fine-Tuning (CFT), which trains models using three types of examples: positive examples that maintain semantic meaning, negative examples with different semantics, and forward-direction obfuscation examples. This approach aims to develop deeper understanding rather than surface-level pattern recognition and allows reverse capabilities to develop naturally without explicit reverse training. Their experiments demonstrated that CFT successfully achieved bidirectional reasoning, enabling strong reverse performance while maintaining forward task capabilities. The authors conclude that bidirectional reasoning serves both as a theoretical framework for assessing genuine understanding and as a practical training approach for developing more capable AI systems.

cs.CL↗

arXiv2025

The Struggles of LLMs in Cross-lingual Code Clone Detection

With the involvement of multiple programming languages in modern software development, cross-lingual code clone detection has gained traction within the software engineering community. Numerous studies have explored this topic, proposing various promising approaches. Inspired by the significant advances in machine learning in recent years, particularly Large Language Models (LLMs), which have demonstrated their ability to tackle various tasks, this paper revisits cross-lingual code clone detection. We evaluate the performance of five (05) LLMs and eight prompts (08) for the identification of cross-lingual code clones. Additionally, we compare these results against two baseline methods. Finally, we evaluate a pre-trained embedding model to assess the effectiveness of the generated representations for classifying clone and non-clone pairs. The studies involving LLMs and Embedding models are evaluated using two widely used cross-lingual datasets, XLCoST and CodeNet. Our results show that LLMs can achieve high F1 scores, up to 0.99, for straightforward programming examples. However, they not only perform less well on programs associated with complex programming challenges but also do not necessarily understand the meaning of "code clones" in a cross-lingual setting. We show that embedding models used to represent code fragments from different programming languages in the same representation space enable the training of a basic classifier that outperforms all LLMs by ~1 and ~20 percentage points on the XLCoST and CodeNet datasets, respectively. This finding suggests that, despite the apparent capabilities of LLMs, embeddings provided by embedding models offer suitable representations to achieve state-of-the-art performance in cross-lingual code clone detection.

cs.SE↗