List your company

Complex Reasoning on BBH (Acc, Time, Peak, Dep)

83.03AccVanilla
5.08225.318545.55565.7915Apr 4, 2026

Evaluation results

MethodLinks
VanillaBackbone=Llama3.1-8B2026.0483.036.82,7935.9
LThinker_thoBackbone=Llama3.1-8B2026.0479.397.469111.9
Vanilla + H2OBackbone=Llama3.1-8B2026.0478.9911.431,0242.1
LThinker_tokBackbone=Llama3.1-8B2026.0474.147.431,5122.9
LThinker_thoBackbone=Qwen2.5-7B2026.0470.37.711,1512.7
Vanilla + H2OBackbone=Qwen2.5-7B2026.0470.115.611,0243.5
VanillaBackbone=Qwen2.5-7B2026.0469.911.53,78613
AnLLMBackbone=Llama3.1-8B2026.0468.899.796211.6
CoTBackbone=Llama3.1-8B2026.0466.460.936420.2
CoTBackbone=Qwen2.5-7B2026.0465.450.685700.1
LThinker_tokBackbone=Qwen2.5-7B2026.0462.838.951,8845.6
Distill-R1Backbone=Llama3.1-8B2026.0461.210.833800.2
Distill-R1Backbone=Qwen2.5-7B2026.0457.785.531,9676
AnLLMBackbone=Qwen2.5-7B2026.0454.9510.041,3033.8
Vanilla + SepLLMBackbone=Llama3.1-8B2026.0414.5527.141,0248.5
Vanilla + SepLLMBackbone=Qwen2.5-7B2026.048.0826.641,0249.4