paper-with-me

홈 › Papers

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

2026-06-23 · Ruslan Berdichevsky, Shai Nahum-Gefen, Elad Ben-Zaken arxiv

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 ($F_1 = 0.305$).

📄 PDF Abstract BibTeX arXiv:2606.25102

Code (0)

등록된 구현이 없습니다.

Tasks

parameter-efficient fine-tuningBinary ClassificationCode Generation

Similar Papers 제목 키워드 기반

SALSA: Single-pass Autoregressive LLM Structured Classification

2025-10-26 · Ruslan Berdichevsky, Shai Nahum-Gefen, Elad Ben Zaken arxiv

Despite their impressive generalization capabilities, instruction-tuned Large Language Models often underperform on text classification benchmarks. We introduce SALSA, a coherent pipeline that combines structured prompti…

parameter-efficient fine-tuningText Classification

SALSA: Sequential Approximate Leverage-Score Algorithm with Application in Analyzing Big Time Series Data

2023-12-30 · Ali Eshragh, Luke Yerbury, Asef Nazari, Fred Roosta 외

We develop a new efficient sequential approximate leverage score algorithm, SALSA, using methods from randomized numerical linear algebra (RandNLA) for large matrices. We demonstrate that, with high probability, the accu…

Time Series

LSALSA: Accelerated Source Separation via Learned Sparse Coding

2018-02-13 · Benjamin Cowen, Apoorva Nandini Saridena, Anna Choromanska

We propose an efficient algorithm for the generalized sparse coding (SC) inference problem. The proposed framework applies to both the single dictionary setting, where each data point is represented as a sparse combinati…

Beyond 1/2-Approximation for Submodular Maximization on Massive Data Streams

2018-07-01 · ICML 2018 7 · Ashkan Norouzi-Fard, Jakub Tarnawski, Slobodan Mitrovic, Amir Zandieh 외

Many tasks in machine learning and data mining, such as data diversification, non-parametric learning, kernel machines, clustering etc., require extracting a small but representative summary from a massive dataset. …

ClusteringRecommendation Systems

Beyond $1/2$-Approximation for Submodular Maximization on Massive Data Streams

2018-08-06 · Ashkan Norouzi-Fard, Jakub Tarnawski, Slobodan Mitrović, Amir Zandieh 외

Many tasks in machine learning and data mining, such as data diversification, non-parametric learning, kernel machines, clustering etc., require extracting a small but representative summary from a massive dataset. Often…

ClusteringRecommendation Systems