List your company
ResearchFormerly WizWand

Commonsense Reasoning on LogiQA

29.8AccuracyDynamic Loss-Based
22.5224.4126.328.19Apr 9, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Dynamic Loss-BasedBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0429.8
DoGraphBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0428.3
DoReMiBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0427.2
DOGEBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0427.2
DoReMiModel=GPT-2 Small, Training Data=SlimPajama2026.0426.8
DoReMiBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0426.4
DoGraphBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0426.3
Dynamic Loss-BasedModel=GPT-2 Small, Training Data=SlimPajama2026.0426
DoGraphModel=GPT-2 Small, Training Data=SlimPajama2026.0426
RegMixBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0425.8
UniformBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0425.7
Data Mixing LawModel=GPT-2 Small, Training Data=SlimPajama2026.0424.5
Data Mixing LawBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0424.4
RegMixModel=GPT-2 Small, Training Data=SlimPajama2026.0424.1
UniformBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0423.9
RegMixBackbone=GPT-2 Medium, Pre-training Dataset=SlimPajama2026.0423.7
Data Mixing LawBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0423.6
DOGEModel=GPT-2 Small, Training Data=SlimPajama2026.0423.5
DOGEBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0423.2
Dynamic Loss-BasedBackbone=LLaMA-1.1B, Pre-training Dataset=The Pile2026.0422.8
UniformModel=GPT-2 Small, Training Data=SlimPajama2026.0422.8