PulseAugur
中
实时 08:55:33
English(EN) DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data

开源 Mimir v1 模型使用道德数据实现前沿性能

研究人员推出了 Mimir v1,这是一个基于分层推理模型 (HRM) 架构构建的 10 亿参数语言模型。该模型在英语方面取得了有竞争力的性能,并在仅使用允许的训练后数据的情况下,在丹麦语方面设定了新的最先进水平。Mimir v1 在 161 个多样化数据集上进行了训练,并在 20 项基准测试中展示了与 Qwen 3.5 4B 和 Gemma 4 E2B 等更大模型相当的性能。 AI

影响 提供了一个具有竞争力的开源替代方案,该方案使用合乎道德来源的数据进行训练,有可能降低研究人员的门槛。

排序理由 关于新语言模型发布的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

开源 Mimir v1 模型使用道德数据实现前沿性能

本文如何被排名

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
关于新语言模型发布的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, paper
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
48 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.AI TIER_1 English(EN) · Peter Schneider-Kamp, Jacob Nielsen, Gianluca Barmina, Kenneth Enevoldsen, Lukas Galke Poech ·

    DFM Mimir v1:一款开放式 HRM,仅使用允许的训练后数据,在 10 亿参数下实现前沿性能

    arXiv:2608.13517v1 Announce Type: cross Abstract: Current large language model development relies on massive, often non-permissible datasets, creating a high barrier for researchers committed to open-source and ethically sourced data. We introduce Mimir v1, a 1-billion-parameter …