LongZero: Enhancing Long-Context Reasoning in LLMs from Zero Data

Hugging Face Daily Papers Papers
long-context-reasoning reinforcement-learning grpo self-improvement llm data-synthesis

Summary

LongZero 让大语言模型自行合成多跳推理与多轮检索任务，并用 GRPO 在零外部数据的条件下自我提升长上下文推理能力，在五个长上下文基准上超越依赖外部语料的基线方法。

Reinforcement learning (RL) is an effective way to strengthen the long-context reasoning of large language models (LLMs), but it needs training data that are long, diverse, challenging, and verifiable. Human annotation of such data is expensive, naturally occurring data are too scarce to sustain RL, and existing synthesis methods depend on external corpora or stronger teacher models, which become harder to obtain as the trained models get stronger. Recently, RL from zero data, in which a model learns from tasks it proposes itself, has achieved excellent performance, but existing methods target short-context tasks. Therefore, we propose LongZero, a method that lets an LLM improve its own long-context reasoning without any external data. Instead of writing questions and answers over existing documents, LongZero first has the model synthesize the question and answer with its supporting evidence in a compact structured form, and only then separately expands the evidence into a long multi-document context. LongZero combines these self-written multi-hop reasoning tasks with self-written multi-round retrieval tasks and trains the model with GRPO. We evaluate LongZero on two LLMs across five long-context reasoning benchmarks and compare it with RL on data synthesized from external corpora. Experimental results show that LongZero can generalize to real tasks, and outperforms both baselines, suggesting a viable and promising path to improving long-context reasoning of LLMs.
Original Article
View Cached Full Text

Cached at: 10/29/26, 01:33 PM

Paper page - LongZero: Enhancing Long-Context Reasoning in LLMs from Zero Data

Source: https://huggingface.co/papers/2610.30132

Abstract

LongZero enables large language models to self-improve long-context reasoning by synthesizing structured multi-hop and retrieval tasks and training with GRPO without external data.

Reinforcement learning(RL) is an effective way to strengthen thelong-context reasoningoflarge language models(LLMs), but it needs training data that are long, diverse, challenging, and verifiable. Human annotation of such data is expensive, naturally occurring data are too scarce to sustain RL, and existing synthesis methods depend on external corpora or stronger teacher models, which become harder to obtain as the trained models get stronger. Recently, RL from zero data, in which a model learns from tasks it proposes itself, has achieved excellent performance, but existing methods target short-context tasks. Therefore, we propose LongZero, a method that lets an LLM improve its ownlong-context reasoningwithout any external data. Instead of writing questions and answers over existing documents, LongZero first has the model synthesize the question and answer with its supporting evidence in a compact structured form, and only then separately expands the evidence into a long multi-document context. LongZero combines these self-writtenmulti-hop reasoningtasks with self-writtenmulti-round retrievaltasks and trains the model withGRPO. We evaluate LongZero on two LLMs across fivelong-context reasoningbenchmarks and compare it with RL on data synthesized from external corpora. Experimental results show that LongZero can generalize to real tasks, and outperforms both baselines, suggesting a viable and promising path to improvinglong-context reasoningof LLMs.

View arXiv pageView PDFGitHub4Add to collection

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2610.30132 in a model README.md to link it from this page.

Datasets citing this paper1

#### zhennan1/LongZero Viewer• Updatedabout 1 hour ago • 7.68k • 73

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2610.30132 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

G-Zero: Self-Play for Open-Ended Generation from Zero Data

Hugging Face Daily Papers

This paper introduces G-Zero, a verifier-free framework that enables autonomous large language model self-improvement through co-evolutionary training using intrinsic rewards and hint-based guidance. It aims to overcome the limitations of proxy LLM judges in open-ended tasks by deriving supervision from internal distributional dynamics.

MindZero: Learning Online Mental Reasoning With Zero Annotations

arXiv cs.AI

MindZero introduces a self-supervised reinforcement learning framework that trains multimodal large language models for efficient and robust online mental reasoning without requiring mental state annotations, outperforming model-based methods in accuracy and efficiency.