List your company

General Reasoning on EvoLM-4B OOD Benchmarks

13.2GPQA ScoreModSFT
3.2165.8088.410.992Jun 7, 2026

Evaluation results

MethodLinks
ModSFTRL=Yes2026.0613.222.96.214.117
RejuvenationRL=Yes2026.0612.926.26.915.317.5
OverSFTRL=Yes2026.069.921.35.612.214.9
DFTRL=Yes2026.068.322.35.712.115.2
OverSFTRL=No2026.067.813.73.38.310
DFTRL=No2026.065.87.23.35.48.5
ModSFTRL=No2026.065.313.43.47.49.1
RejuvenationRL=No2026.063.62.62.12.85.4