AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

GEM: Geometric Entropy Mixing for Optimal LLM Data Curation

Yue Min, Ziyun Qiao, Ruining Chen, Yujun Li

ICML 2026 · Large language models › Language, safety and interpretability · approach: LLM pruning and scaling data · 0 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
add vMF mixture with soft assignments instead of spherical K-Means (ablation)GEM clustering · Clustering ablationaverage accuracy: 40.6 -> 41.3Improved
add mixing-balance regularizer and entropy constraint to vanilla vMF (ablation)GEM clustering · Clustering ablationaverage accuracy: 41.3 -> 42.1Improved
use spherical cosine geometry instead of Euclidean K-Means (ablation)clustering · Clustering ablationaverage accuracy: 38.5 -> 40.6Improved
use GIS instead of random sampling for prototype selection (ablation)taxonomy generation · held-out labelingaccuracy: 80.41 -> 84.17Improved
use GIS instead of center-only selection (ablation)taxonomy generation · held-out labelingaccuracy: 81.67 -> 84.17Improved
use GIS instead of confidence-only selection (ablation)taxonomy generation · held-out labelingaccuracy: 80.83 -> 84.17Improved
raise K from 24 to 36 (ablation)GEM + Perf · downstream averageaccuracy: 41.21%Improved
raise K from 36 to 48 (ablation)GEM + Perf · downstream averageaccuracyHurt
use GEM taxonomy instead of WebOrganizer under DoReMiDoReMi mixing · downstream averageaverage accuracy: 42.79 -> 43.95Improved
use GEM taxonomy instead of WebOrganizer under PerfPerf mixing · downstream averageaverage accuracy: 44.25 -> 44.79Improved
use GEM taxonomy instead of WebOrganizer under RegMixRegMix mixing · downstream averageaverage accuracy: 40.77 -> 41.45Improved
use GEM partitions instead of K-Means to distill FastText studentstudent classifier · test splitaccuracy: 72.92 -> 75.13Improved

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