English(EN) Efficient LLM Adversarial Training via Low-Rank Defense and Circuit-Guided Surrogates

新方法大幅降低 LLM 对抗训练成本

作者 PulseAugur 编辑部 · [1 个来源] · 2026-08-03 04:00

研究人员开发了一种计算效率更高的大型语言模型（LLM）对抗训练方法。这种新方法优化了训练过程的防御和攻击双方。在防御方面，它利用表示微调（ReFT）并解决了潜在的 token 选择问题。在攻击方面，它通过仅提取 LLM 的相关电路来构建轻量级代理模型，与完整模型微调相比，显著降低了计算成本。 AI

影响降低了 LLM 对抗训练的计算成本，可能使更强大的 LLM 开发更易于获得。

排序理由关于 LLM 对抗训练新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

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

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 LLM 对抗训练新方法的学术论文。[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: paper, infra; 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: 85 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。

arXiv cs.AI TIER_1 English(EN) · Weiyi He, Yuping Lin, Jiliang Tang, Yue Xing · 2026-08-03 04:00

通过低秩防御和电路引导的代理实现高效 LLM 对抗性训练

arXiv:2607.28959v1 Announce Type: cross Abstract: Adversarial training is one of the most effective defenses against adversarial attacks, yet the computational cost remains prohibitive at modern scales, especially for large language models (LLMs). While existing mitigation strate…