arXiv cs.CL· Xinyu Zhu, Yihao Feng, Yanchao Sun, Xianzhi Du, Pingzhi Li, Olli Saarikivi, Yun Zhu, Yu Meng·· 12 小时前AI 评分38
CHIMERA：面向通用 LLM 推理的紧凑合成数据集
CHIMERA: Compact Synthetic Data for Generalizable LLM Reasoning
AI 导读
研究团队推出 CHIMERA，一个仅 9K 样本的紧凑合成推理数据集，覆盖 8 个科学学科、超 1K 个细粒度主题，并提供由强推理模型合成的长 CoT 轨迹与全自动交叉验证评估流程。
正文 · 原文
Abstract:Large Language Models (LLMs) have exhibited remarkable reasoning capabilities, largely enabled by supervised fine-tuning (SFT)- and reinforcement learning (RL)-based post-training on high-quality reasoning data. However, reproducing and extending these capabilities in open and scalable settings is hindered by three data-centric challenges: (1) the cold-start problem, arising from the lack of seed datasets with detailed, long Chain-of-Thought (CoT) trajectories needed to initialize reasoning policies; (2) limited domain coverage, as most existing open-source reasoning datasets are concentrated in mathematics, with limited coverage of broader scientific disciplines; and (3) the annotation bottleneck, where the difficulty of frontier-level reasoning tasks makes reliable human annotation prohibitively expensive or infeasible. To address these challenges, we introduce CHIMERA, a compact synthetic reasoning dataset of 9K samples designed to support generalizable reasoning across domains. CHIMERA is constructed with three key properties: (1) it provides rich, long CoT reasoning trajectories synthesized by state-of-the-art reasoning models; (2) it has broad and structured coverage, spanning 8 major scientific disciplines and over 1K fine-grained topics organized via a model-generated hierarchical taxonomy; and (3) it employs a fully automated, scalable evaluation pipeline that uses strong reasoning models to cross-validate both problem validity and answer correctness, removing the reliance on human annotations. We use CHIMERA to post-train a 4B Qwen3 model with SFT and RL. Despite the dataset's modest size, the resulting model achieves strong performance on a suite of challenging reasoning benchmarks, including GPQA-Diamond, AIME 24/25/26, HMMT 25 and Humanity's Last Exam, approaching or matching the performance of substantially larger models such as DeepSeek-R1 and Qwen3-235B.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2603.00889 [cs.CL] |
| (or arXiv:2603.00889v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2603.00889 arXiv-issued DOI via DataCite |
Submission history
From: Xinyu Zhu [view email]
[v1]
Sun, 1 Mar 2026 03:23:41 UTC (832 KB)
[v2]
Tue, 20 Oct 2026 05:13:22 UTC (873 KB)
来源：arXiv cs.CL · arxiv.org