AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Rank-Aware Differentially Private Release of Listwise Preferences for LLM Alignment

Junwei Chen, Manjiang Yu, Pengpeng Qiao, Yang Cao

NeurIPS 2026 · Federated learning and privacy › Differential privacy · approach: Differentially private synthetic data · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a rank-aware local differential privacy mechanism for releasing listwise preference feedback to align LLMs, extending privacy-preserving alignment beyond pairwise comparisons.

Problem

Listwise preference feedback reveals more sensitive user preferences than pairwise comparisons, and privacy-preserving alignment under local differential privacy has mainly addressed pairwise feedback, leaving listwise alignment under LDP unexplored.

Approach

Extends local differential privacy to listwise preference release by making the privatization mechanism rank-aware, so that privacy noise is tailored to the structure of full rankings rather than treating them as independent pairwise comparisons.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace randomized response with rank-aware exponential mechanism for private binary releaseKTO on privatized listwise labels (Qwen2-0.5B) · ListUltraFeedback, eps=0.5average win rate vs SFT: 48.0 -> 51.2Improved
replace randomized response with rank-aware exponential mechanismKTO (Qwen2-0.5B) · ListUltraFeedback, eps=3average win rate vs SFT: 57.1 -> 59.6Improved
replace randomized response with rank-aware exponential mechanismKTO (Mistral-7B) · ListUltraFeedback, eps=0.5average win rate vs SFT: 84.0 -> 85.2Improved
replace randomized response with rank-aware exponential mechanismKTO (Mistral-7B) · ListUltraFeedback, eps=3average win rate vs SFT: 85.8 -> 86.7Improved
replace randomized response with rank-aware exponential mechanismKTO (Qwen2-0.5B) · MT-Bench, eps=0.5win rate vs SFT: 52.5 -> 50.1Hurt
replace randomized response with rank-aware exponential mechanismKTO (Qwen2-0.5B) · AlpacaEval, eps=1win rate vs SFT: 50.7 -> 50.6No clear effect
replace randomized response with rank-aware exponential mechanismKTO (Qwen2-0.5B) · Scoring (ArmoRM) win rate, eps=1win rate vs SFT: 43.7 -> 49.1Improved
release binary top-B labels instead of cardinal scores or full ranks to cut sensitivityprivate release of listwise preferences · sensitivity analysis (m items)sensitivity / noise scale: m^2 -> 2(m-1) -> 2Improved
use linear boundary score rather than other monotone rank scores (ablation)rank-aware utility · calibrated rank span (theory)S_w/Delta_wImproved
add privatization at private privacy budget epsilon compared with non-private trainingKTO (Qwen2-0.5B) · ListUltraFeedback average, eps=3 vs non-privateaverage win rate vs SFT: 60.8 -> 59.6Hurt
keep full candidate list for private release instead of random small subset (ablation)RAEM+KTO (Qwen2-0.5B) · ListUltraFeedback, eps=1average proxy win rateImproved
increase list size m under oracle top-bottom selection (ablation)RAEM+KTO (Qwen2-0.5B) · ListUltraFeedback, eps=1average proxy win rateNo clear effect

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