Research map / Large language models / Alignment and preferences
Cultivating Pluralism In Algorithmic Monoculture: The Community Alignment Dataset
Lily H Zhang, Smitha Milli, Karen Long Jusko, Jonathan Smith, Brandon Amos, Wassim Bouaziz, Manon Revel, Jack Kussman, Yasha Sheynin, Lisa Titus, Bhaktipriya Radharapu, Jane Yu and 3 more
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 |
|---|---|---|---|
| use negatively-correlated sampling instead of temperature sampling for candidate responses | candidate response sets (traditional values) · everyday chatbot prompts, 4 candidates | mean coverage of traditional values: 15% -> 60% | Improved |
| use negatively-correlated sampling instead of temperature sampling for candidate responses | candidate response sets (survival values) · everyday chatbot prompts, 4 candidates | mean coverage of survival values: 30% -> 53% | Improved |
| build preference data with NC sampling instead of temperature sampling (PRISM prompts) | SFT, DPO, GRPO and prompt steering on Llama 3.1-8B and 3.3-70B · PRISM prompts, IW value win rate | win rate vs original model: near chance -> around 70-90% | Improved |
| sample candidates from 21 LLMs instead of 1 LLM at temperature 1 (ablation) | preference dataset for alignment methods · PRISM prompts, IW values | win rate vs original model | No clear effect |
| raise sampling temperature (0, 0.5, 1) (ablation) | LLM response generation · IW value coverage | coverage of IW dimensions | No clear effect |
Closest papers by meaning
- Group Preference Optimization: Few-Shot Alignment of Large Language Models (ICLR 2024 · similarity 0.91)
- Latent Preference Coding: Aligning Large Language Models via Discrete Latent Codes (ICML 2025 · similarity 0.90)
- The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Models (NeurIPS 2024 · similarity 0.90)
- Aligning to Thousands of Preferences via System Message Generalization (NeurIPS 2024 · similarity 0.89)
- Understanding the Learning Dynamics of Alignment with Human Feedback (ICML 2024 · similarity 0.89)
- Can DPO Learn Diverse Human Values? A Theoretical Scaling Law (NeurIPS 2025 · 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.
