List your company

STEM Reasoning on Minerva (Accuracy)

32AccuracyGeoMin
23.99226.07128.1530.229Jun 3, 2026

Evaluation results

MethodLinks
GeoMinBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0632
Seq-entropyBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0627.6
Self-certaintyBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0627.6
Tok-entropyBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0627.2
TTRLBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0625.7
Co-rewardingBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0624.6
TraPOBackbone=Deepseek-R1-Distill-Qwen-1.5B, Supervision Setting=semi-supervised (10% labeled data), Domain Context=In-Domain2026.0624.3