Research map / Large language models / Alignment and preferences
Principle-Driven Self-Alignment of Language Models from Scratch with Minimal Human Supervision
Zhiqing Sun, Yikang Shen, Qinhong Zhou, Hongxin Zhang, Zhenfang Chen, David Daniel Cox, Yiming Yang, Chuang Gan
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add verbose cloning after principle engraving (ablation) | Dromedary-65B · HHH Eval | MC accuracy (Overall): 0.85 -> 0.83 | Hurt |
| add verbose cloning after principle engraving (ablation) | Dromedary-65B · HHH Eval Other | MC accuracy: 0.88 -> 0.81 | Hurt |
| add verbose cloning after principle engraving (ablation) | Dromedary-65B · Vicuna benchmark questions (GPT-4 judged) | response quality | Improved |
| fine-tune on self-aligned outputs (principle engraving) instead of prompting the base model with principles (ablation) | LLaMA-65B · alignment benchmarks | alignment performance | Improved |
Closest papers by meaning
- The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning (ICLR 2024 · similarity 0.93)
- Linear Alignment: A Closed-form Solution for Aligning Human Preferences without Tuning and Feedback (ICML 2024 · similarity 0.90)
- SALMON: Self-Alignment with Instructable Reward Models (ICLR 2024 · similarity 0.90)
- Preference Ranking Optimization for Human Alignment (AAAI 2024 · similarity 0.90)
- RLTHF: Targeted Human Feedback for LLM Alignment (ICML 2025 · similarity 0.89)
- Self-Supervised Alignment with Mutual Information: Learning to Follow Principles without Preference Labels (NeurIPS 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
