List your company

Science Question Answering on GPQA (Accuracy, Memory, and Time)

0.21Memory RatioAC
0.17840.39170.6050.8183Jan 29, 2026

Evaluation results

MethodLinks
ACBackbone model=Qwen3-4B2026.010.212.1165.03
T2Backbone model=Llama3.1-8B2026.010.250.3736.5
T2Backbone model=Deepseek-R1-Distill-Qwen-7B2026.010.250.4555.53
DSCBackbone model=Llama3.1-8B2026.010.262.0835.92
T2Backbone model=Qwen2.5-7B2026.010.270.5139.17
T2Backbone model=Qwen3-4B2026.010.270.4665.17
DSCBackbone model=Qwen2.5-7B2026.010.312.1838.12
ESCBackbone model=Llama3.1-8B2026.010.322.3836.45
ESCBackbone model=Qwen3-4B2026.010.321.9765.12
ESCBackbone model=Qwen2.5-7B2026.010.362.2739.05
DSCBackbone model=Qwen3-4B2026.010.363.9264.21
ACBackbone model=Llama3.1-8B2026.010.372.9936.24
ACBackbone model=Qwen2.5-7B2026.010.382.738.79
ESCBackbone model=Deepseek-R1-Distill-Qwen-7B2026.010.431.6755.42
DeepConfBackbone model=Qwen3-4B2026.010.493.264.87
DeepConfBackbone model=Llama3.1-8B2026.010.524.1734.49
DeepConfBackbone model=Qwen2.5-7B2026.010.533.4139.19
ACBackbone model=Deepseek-R1-Distill-Qwen-7B2026.010.542.6254.67
DSCBackbone model=Deepseek-R1-Distill-Qwen-7B2026.010.64.0352.99
DeepConfBackbone model=Deepseek-R1-Distill-Qwen-7B2026.010.653.853.13
Std-PT.Backbone model=Qwen2.5-7B2026.011139.49
Std-PT.Backbone model=Llama3.1-8B2026.011136.66
Std-PT.Backbone model=Deepseek-R1-Distill-Qwen-7B2026.011155.29
Std-PT.Backbone model=Qwen3-4B2026.011165.33