The paper introduces LM2LM, a method that trains large language models (LLMs) to propose improvements for other LLMs using reinforcement learning (RL). It employs a frozen research world model (RWM) to predict the gains of proposals, providing a cheaper surrogate reward, and periodically validates high‑scoring proposals to keep feedback reliable. Experiments across eight LM development environments show that LM2LM enables a Qwen3.8‑27B model to outperform GPT‑6‑sol on held‑out tasks while reducing data‑generation cost by 33%, and it generalizes to unseen model families.
By Minhua Lin, Zijun Wang, Zewen Liu, Zhaotian Weng, Bing He, Zhan Shi, Benoit Dumoulin, Xiang Zhang, Suhang Wang, Hanqing Lu
arXiv:2609.39714v1 Announce Type: new
Abstract: Top researchers have good intuition, but do language models have as good intuition about model training as top AI researchers? To measure model intuiti...
By Zirui Ren, Shaoyang Guo, Chencheng Tang, Jinxin Wang, Chengyu Xiong, Shanbin Yu, Peihang Li, Yidi Wu, Bangzhe Huang, Qingyu Qu, Leqian Yang, Ziming Liu
The paper proposes using language models as Research World Models (RWMs) to predict outcomes of experimental interventions across diverse research environments. By training on over 2,600 experimental records and 171,000 GPU‑hours, RWMs improve prediction accuracy within and across environments, reducing selection regret by up to 78% and boosting gains in multi‑round autoresearch. Ablation studies show that incorporating research knowledge yields greater benefits than merely changing the language model or increasing reasoning effort.
By Zijun Wang, Zewen Liu, Minhua Lin, Zhaotian Weng, Zhan Shi, Bing He, Yisi Sang, Dakuo Wang, Benoit Dumoulin, Wei Jin, Yuyin Zhou, Cihang Xie, Hanqing Lu
DataSense-Bench is a new benchmark that tests whether advanced AI models can choose and rank training data subsets that will improve a small language model’s performance. The benchmark lets agents inspect data, run analysis code, and perform forward passes, but not train the model or access evaluation tasks. After agents select subsets, the base model is fine‑tuned on each subset and evaluated, measuring both the performance of the best subset and the accuracy of the agents’ rankings. Experiments show limited gains over random selection and inconsistent ranking ability across tasks, with agents often using similar data signals but interpreting their value differently.
By Yudi Zhang, Mingyu Cao, Lu Yin, Mykola Pechenizkiy, Shiwei Liu
arXiv:2607. 14109v1 Announce Type: cross Abstract: Probing the capabilities of Large Language Models (LLMs) and building robust solutions for Multiple-Choice Question Answering (MCQA) remain central challenges in natural language understanding.
By Inder Preet, Shuxin Lin, Dhaval Patel
MLExpert is a 35B Mixture-of-Experts agentic model designed to autonomously conduct long‑horizon machine learning engineering (MLE) tasks. It uses a post‑training approach that combines hierarchical supervised fine‑tuning with Windowed Interaction RL to learn from intermediate outcomes in complex experimentation workflows. On the MLE‑Bench Lite benchmark, MLExpert raises the average medal rate from 37.9% to 72.7%, matching the performance of larger frontier models and demonstrating transferable optimization capabilities to other benchmarks such as NatureBench Lite.