Published August 24, 2026 | Version v1.1.0

Evaluator software genealogy exposes hidden dependence in world-model benchmarks

  • 1. Harbin Institute of Technology
  • 2. University of Belgrade, Faculty of Organizational Sciences

Description

Reproducibility archive for the study 'Evaluator software genealogy exposes hidden dependence in world-model benchmarks'. It contains source-verifiable evaluator-lineage records, derived analysis inputs and outputs, conclusion-robustness certificates, source data, figure-generation code, tests, and verification utilities.

Files

cuijianxun/evaluator-lineages-v1.1.0.zip

Files (3.6 MB)

Name Size
md5:da178dafd9e40ae16337e3021a8975ba
3.6 MB Preview Download

Additional details