提出逆语言建模框架，让大模型更鲁棒且能识别有害输入触发词。
Inverse Language Modeling towards Robust and Grounded LLMs
- 逆向生成输出，反推潜在有害输入触发词
- 同时提升模型对输入扰动的鲁棒性
- 适合安全测试与可控大模型研发
当前大模型防御机制分散且发展不足，远不及分类器研究。为提升大模型的对抗鲁棒性，我们提出逆语言建模（ILM），一种统一框架，可同时实现：1）增强大模型对输入扰动的鲁棒性；2）通过反转模型输出，识别潜在有毒或不安全的输入触发词，实现原生具身化。ILM将大模型从静态生成器转变为可分析、鲁棒的系统，有望助力红队测试。该方法或可奠定下一代既鲁棒又具身、且更可控可信的大模型基础。代码已公开于 github.com/davegabe/pag-llm。
原文摘要 · Abstract (English)
The current landscape of defensive mechanisms for LLMs is fragmented and underdeveloped, unlike prior work on classifiers. To further promote adversarial robustness in LLMs, we propose Inverse Language Modeling (ILM), a unified framework that simultaneously 1) improves the robustness of LLMs to input perturbations, and, at the same time, 2) enables native grounding by inverting model outputs to identify potentially toxic or unsafe input triggers. ILM transforms LLMs from static generators into analyzable and robust systems, potentially helping RED teaming. ILM can lay the foundation for next-generation LLMs that are not only robust and grounded but also fundamentally more controllable and trustworthy. The code is publicly available at github.com/davegabe/pag-llm.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。