AI Paper Atlas

Research map / Large language models / Alignment and preferences

Operationalizing Pluralistic Values in Large Language Model Alignment Reveals Trade-offs in Safety, Inclusivity, and Model Behavior

Dalia Ali, Dora Zhao, Allison Koenecke, Orestis Papakyriakopoulos

AAAI 2026 · Large language models › Alignment and preferences · approach: Robust and calibrated DPO · 15 citations (OpenAlex)

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 changedWhereMeasuredEffect
use 5-point Likert ratings instead of binary (ablation)DPO toxicity fine-tuning · Toxicitytoxicity effect size: -0.198 -> -0.242Improved
use 5-point Likert ratings instead of 3-point (ablation)DPO toxicity fine-tuning · Toxicitytoxicity effect size: -0.225 -> -0.242Improved
preserve all ratings instead of majority vote (ablation)DPO toxicity fine-tuning · Toxicitytoxicity effect size: -0.158 -> -0.242Improved
preserve all ratings instead of full consensus (ablation)DPO toxicity fine-tuning · Toxicitytoxicity effect size: -0.039 -> -0.242Improved
preserve all ratings instead of random selection (ablation)DPO toxicity fine-tuning · Toxicitytoxicity effect size: -0.146 -> -0.242Improved
preserve all ratings instead of averaging (ablation)DPO toxicity fine-tuning · Toxicitytoxicity effect size: -0.229 -> -0.242No clear effect
use DPO instead of GRPO for multi-objective alignmentToxic+EA training · Toxicitytoxicity effect size: -0.020 -> -0.159Improved
use DPO instead of GRPO for multi-objective alignmentToxic+EA training · Emotional awarenessEA effect size: 0.029 -> 0.084Improved
single-objective toxicity DPO instead of multi-objective Toxic+EA (ablation)DPO · Toxicitytoxicity effect size: -0.159 -> -0.243Improved
single-objective EA DPO instead of multi-objective Toxic+EA (ablation)DPO · Emotional awarenessEA effect size: 0.084 -> 0.083No clear effect
fine-tune on Liberal instead of Conservative feedbackDPO · Emotional awarenesspooled effect: 0.049Improved
fine-tune on White instead of Black feedbackDPO · Emotional awarenesspooled effect: 0.046Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy