titleIncentivizing Self-Learning in Large Language Models via Reinforcement Learning
authorsLingkai Kong, Yuzhe Gu, Haiteng Zhao, Zijian Wu, Youqing Fang, Zhouqi Hua, Yanan Sun, Kai Chen, Wenwei Zhang
date2026-10-29
venuearXiv (preprint, not peer reviewed)
idarXiv:2610.39671
affiliationIntern Large Models
upvotes31 (Hugging Face Daily Papers, 2026-11-14)
linksabs pdf hf
cite
relatedAtria Dawn: The Dawn of Agentic Superintelligence
Sage S2.1 Flash Technical Report
Intern-S2-Preview: Scientific Agentic Foundation Model
Scalable Visual Pretraining for Language Intelligence
Beyond the Current Observation: Evaluating Multimodal Large Language Models in Controllable Non-Markov Games
abstractLarge language models (LLMs) draw on parametric knowledge and retrieval-augmented generation (RAG) to solve knowledge-intensive tasks. Yet problem solving in unfamiliar domains calls for tightly coupled knowledge acquisition and reasoning, beyond factual retrieval and aggregation or learning solely from a preassembled context. We study self-learning -- autonomously retrieving, studying, and applying knowledge from long-form reference materials on the fly. To investigate this capability, we formulate a textbook-based mathematical problem-solving task. LLMs must navigate textbook exposition, learn and apply relevant principles, and construct proofs for exercises, with reference solutions withheld. We introduce a BookNavigator interface to support this human-inspired study process and incentivize self-learning through reinforcement learning (RL) with fine-grained rewards. Experiments show that our method improves performance across in-domain and out-of-domain textbooks and subjects, raising Intern-S2-Preview-35B's overall proof accuracy from 51.5% to 79.5% and outperforming RL-trained textbook-free and fixed-retrieval baselines. These gains extend to complex long-context benchmarks involving different task types, with improvements over the base model of 1.7 and 2.3 percentage points on BrowseComp-Plus and CL-bench, respectively. Together, these findings suggest that RL can foster a domain-agnostic self-learning strategy, highlighting the potential of self-learning from long-form corpora to strengthen broader problem-solving capabilities on complex, long-context tasks.

bibtex

@misc{kong2026incentivizing,
  title={Incentivizing Self-Learning in Large Language Models via Reinforcement Learning},
  author={Lingkai Kong and Yuzhe Gu and Haiteng Zhao and Zijian Wu and Youqing Fang and Zhouqi Hua and Yanan Sun and Kai Chen and Wenwei Zhang},
  year={2026},
  eprint={2610.39671},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2610.39671}
}