Research map / Large language models / Language, safety and interpretability
Beyond Exchangeability: The Chinese Voting Process
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Replace CRP vote-count selection with CVP display-rank selection | CVP selection phase · StackExchange and Amazon, first 50 actions (average) | selection negative log-likelihood: 2.050 -> 1.945 | Helped |
| Add intrinsic response quality q (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .103 -> .095 | Helped |
| Add opinion-ratio feature lambda (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .109 -> .095 | Helped |
| Add item-specific length bias nu_i (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .099 -> .095 | Helped |
| Add lambda and nu_i to q (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .109 -> .095 | Helped |
| Add lambda to q and nu_i (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .105 -> .095 | Helped |
| Add q and nu_i to lambda (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .103 -> .095 | Helped |
| Add q to lambda and nu_i (ablation) | CVP voting model · StackExchange and Amazon, first 50 actions (average) | voting negative log-likelihood: .098 -> .095 | Helped |
| Rank responses by inferred intrinsic quality rather than display rank | CVP quality analysis · StackExchange and Amazon, final snapshot (average) | residual against comment sentiment: .011 -> .006 | Helped |
| Rank responses by inferred intrinsic quality rather than display rank | CVP quality analysis · StackExchange and Amazon, final snapshot (average) | bumpiness against comment sentiment: .186 -> .101 | Helped |
Closest papers by meaning
- Counterfactual Voting Adjustment for Quality Assessment and Fairer Voting in Online Platforms with Helpfulness Evaluation (ICML 2025 · similarity 0.85)
- Computing Voting Rules with Improvement Feedback (ICML 2025 · similarity 0.81)
- Reaching a Consensus in Predictive Loops (NeurIPS 2026 · similarity 0.80)
- Representative Ranking for Deliberation in the Public Sphere (ICML 2025 · similarity 0.80)
- Generative Social Choice: The Next Generation (ICML 2025 · similarity 0.80)
- AI-Mediated Communication Can Steer Collective Opinion (NeurIPS 2026 · similarity 0.80)
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.