[2406.08464] Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. arXiv Operational Status Get status notifications via email or slack
[2406.08464] Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing Skip to main content arXiv is now an independent nonprofit! Learn more × Search arXiv Press Enter to search · Advanced search --> Computer Science > Computation and Language arXiv:2406.08464 (cs) [Submitted on 12 Jun 2024 ( v1 ), last revised 7 Oct 2024 (this version, v2)] Title: Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing Authors: Zhangchen Xu , Fengqing Jiang , Luyao Niu , Yuntian Deng , Radha Poovendran , Yejin Choi , Bill Yuchen Lin View a
Explore this link on the map →related reading
- Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment — LessWronglesswrong.com
- GenAI Handbookgenai-handbook.github.io
- Teaching Claude why \ Anthropicanthropic.com
- [2203.02155] Training language models to follow instructions with human feedbackarxiv.org
- Subliminal Learning: Language Models Transmit Behavioral Traits via Hidden Signals in Dataalignment.anthropic.com
- [2502.17424] Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMsarxiv.org
- GitHub - jacobhilton/deep_learning_curriculum: Language model alignment-focused deep learning curriculum · GitHubgithub.com
- Self-Fulfilling Misalignment Data Might Be Poisoning Our AI Modelsturntrout.com
- How to Generate and Use Synthetic Data for Finetuningeugeneyan.com
- Automated Alignment Researchers: Using large language models to scale scalable oversight \ Anthropicanthropic.com
- Unsupervised Elicitationalignment.anthropic.com
- Stanford CRFMcrfm.stanford.edu