Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Zongqian, Lv, Tengchao, Huang, Shaohan, Su, Yixuan, Sun, Qinzheng, Yin, Qiufeng, Xin, Ying, Li, Scarlett, Cui, Lei, Collier, Nigel, Wei, Furu
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2603.07779
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866914379552784384
author Li, Zongqian
Lv, Tengchao
Huang, Shaohan
Su, Yixuan
Sun, Qinzheng
Yin, Qiufeng
Xin, Ying
Li, Scarlett
Cui, Lei
Collier, Nigel
Wei, Furu
author_facet Li, Zongqian
Lv, Tengchao
Huang, Shaohan
Su, Yixuan
Sun, Qinzheng
Yin, Qiufeng
Xin, Ying
Li, Scarlett
Cui, Lei
Collier, Nigel
Wei, Furu
contents Training next-generation code generation models requires high-quality datasets, yet existing datasets face difficulty imbalance, format inconsistency, and data quality problems. We address these challenges through systematic data processing and difficulty scaling. We introduce a four-stage Data Processing Framework encompassing collection, processing, filtering, and verification, incorporating Automatic Difficulty Filtering via an LLM-based predict-calibrate-select framework that leverages multi-dimensional difficulty metrics across five weighted dimensions to retain challenging problems while removing simplistic ones. The resulting MicroCoder dataset comprises tens of thousands of curated real competitive programming problems from diverse platforms, emphasizing recency and difficulty. Evaluations on strictly unseen LiveCodeBench demonstrate that MicroCoder achieves 3x larger performance gains within 300 training steps compared to widely-used baseline datasets of comparable size, with consistent advantages under both GRPO and its variant training algorithms. The MicroCoder dataset delivers obvious improvements on medium and hard problems across different model sizes, achieving up to 17.2% relative gains in overall performance where model capabilities are most stretched. These results validate that difficulty-aware data curation improves model performance on challenging tasks, providing multiple insights for dataset creation in code generation.
format Preprint
id arxiv_https___arxiv_org_abs_2603_07779
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Scaling Data Difficulty: Improving Coding Models via Reinforcement Learning on Fresh and Challenging Problems
Li, Zongqian
Lv, Tengchao
Huang, Shaohan
Su, Yixuan
Sun, Qinzheng
Yin, Qiufeng
Xin, Ying
Li, Scarlett
Cui, Lei
Collier, Nigel
Wei, Furu
Computation and Language
General Literature
Machine Learning
Training next-generation code generation models requires high-quality datasets, yet existing datasets face difficulty imbalance, format inconsistency, and data quality problems. We address these challenges through systematic data processing and difficulty scaling. We introduce a four-stage Data Processing Framework encompassing collection, processing, filtering, and verification, incorporating Automatic Difficulty Filtering via an LLM-based predict-calibrate-select framework that leverages multi-dimensional difficulty metrics across five weighted dimensions to retain challenging problems while removing simplistic ones. The resulting MicroCoder dataset comprises tens of thousands of curated real competitive programming problems from diverse platforms, emphasizing recency and difficulty. Evaluations on strictly unseen LiveCodeBench demonstrate that MicroCoder achieves 3x larger performance gains within 300 training steps compared to widely-used baseline datasets of comparable size, with consistent advantages under both GRPO and its variant training algorithms. The MicroCoder dataset delivers obvious improvements on medium and hard problems across different model sizes, achieving up to 17.2% relative gains in overall performance where model capabilities are most stretched. These results validate that difficulty-aware data curation improves model performance on challenging tasks, providing multiple insights for dataset creation in code generation.
title Scaling Data Difficulty: Improving Coding Models via Reinforcement Learning on Fresh and Challenging Problems
topic Computation and Language
General Literature
Machine Learning
url https://arxiv.org/abs/2603.07779