List your company

Dataset Diversity Analysis on Post-training LLM Datasets

452.44Vendi ScoreOurs (Provenance-based)
95.148187.9065280.665373.4235Apr 12, 2026

Evaluation results

MethodLinks
Ours (Provenance-based)Size=570K2026.04452.440.6385
OpenHermes-2.5Size=615K2026.04437.760.6271
herculesv1Size=463K2026.04397.330.6121
tulu-3-sft-mixtureSize=939K2026.04375.780.6169
MegaScienceSize=1.2M2026.04373.780.615
TextbookReasoningSize=651K2026.04283.750.5598
OmniThought-0528Size=301K2026.04162.520.514
OpenThoughts3Size=1.2M2026.04133.260.497
MiroMind-M1-SFTSize=719K2026.04108.890.4597