List your company

Code Generation on HumanEval (Accuracy Metric)

96.3HumanEval AccuracyWeak-to-Strong (AEA)
-3.85222.14948.1574.151Apr 1, 2026Apr 9, 2026Apr 18, 2026Apr 26, 2026May 5, 2026May 13, 2026May 22, 2026

Evaluation results

MethodLinks
Weak-to-Strong (AEA)Base Model=GPT-4o, Aligned=true, AEA Model=AEA-4B2026.0596.3
Single AgentBase Model=GPT-4o, Aligned=false, AEA Model=None2026.0595.7
Self-Reflection (AEA)Base Model=GPT-4o, Aligned=true, AEA Model=GPT-4o2026.0594.5
Self-Reflection (AEA)Base Model=GPT-4o-mini, Aligned=true, AEA Model=GPT-4o-mini2026.0593.3
Weak-to-Strong (AEA)Base Model=GPT-4o-mini, Aligned=true, AEA Model=AEA-4B2026.0592.2
Multi AgentBase Model=GPT-4o-mini, Aligned=false, AEA Model=None2026.0592
Single AgentBase Model=GPT-4o-mini, Aligned=false, AEA Model=None2026.0591.5
Multi AgentBase Model=GPT-4o, Aligned=false, AEA Model=None2026.0591.5
Weak-to-Strong (AEA)Base Model=Claude 3.7 Sonnet, Aligned=true, AEA Model=AEA-4B2026.0590.9
Self-Reflection (AEA)Base Model=Claude 3.7 Sonnet, Aligned=true, AEA Model=Claude 3.7 Sonnet2026.0589.6
Self-Reflection (AEA)Base Model=Claude 3.5 Sonnet, Aligned=true, AEA Model=Claude 3.5 Sonnet2026.0589
Weak-to-Strong (AEA)Base Model=Claude 3.5 Sonnet, Aligned=true, AEA Model=AEA-4B2026.0586.6
Multi AgentBase Model=Claude 3.7 Sonnet, Aligned=false, AEA Model=None2026.0585.4
Multi AgentBase Model=Claude 3.5 Sonnet, Aligned=false, AEA Model=None2026.0584.8
Single AgentBase Model=Claude 3.7 Sonnet, Aligned=false, AEA Model=None2026.0584.8
Single AgentBase Model=Claude 3.5 Sonnet, Aligned=false, AEA Model=None2026.0582.3
Multi AgentBase Model=Claude 3 Haiku, Aligned=false, AEA Model=None2026.0579.8
Weak-to-Strong (AEA)Base Model=Claude 3 Haiku, Aligned=true, AEA Model=AEA-4B2026.0579.3
SFTModel=DS-R1-Qwen-14B2026.0477.4
D-QRELOModel=DS-R1-Qwen-14B2026.0477.4
WandaModel=DS-R1-Qwen-14B2026.0475.8
MagnitudeModel=DS-R1-Qwen-14B2026.0475.2
BitDeltaModel=DS-R1-Qwen-14B2026.0473.8
BackboneModel=DS-R1-Qwen-14B2026.0472.6
Self-Reflection (AEA)Base Model=Claude 3 Haiku, Aligned=true, AEA Model=Claude 3 Haiku2026.0572
RandomModel=DS-R1-Qwen-14B2026.0469.5
SVDModel=DS-R1-Qwen-14B2026.0469.5
MoBA (k=2)w_eff=1282026.0468.3
MoBA (k=2)w_eff=2562026.0467.1
Delta-CoMeModel=DS-R1-Qwen-14B2026.0466.5
ISTBase Model=Qwen3-8B, Params=334M2026.0565.9
Stochasticw_eff=2562026.0465.2
DomLoRABase Model=Qwen3-8B, Params=2.1M2026.0564.6
Fullw_eff=∞2026.0464
Stochasticw_eff=1282026.0464
PLoPBase Model=Qwen3-8B, Params=131M2026.0562.8
Single AgentBase Model=Claude 3 Haiku, Aligned=false, AEA Model=None2026.0561.6
LoRABase Model=Qwen3-8B, Params=334M2026.0561
MoBA (k=2)w_eff=642026.0460.4
DomLoRABase Model=LLaMA-3.1-8B, Params=2.3M2026.0553
Stochasticw_eff=642026.0452.4
SWAw_eff=2562026.0452.4
LoRABase Model=LLaMA-3.1-8B, Params=321M2026.0545.1
ISTBase Model=LLaMA-3.1-8B, Params=321M2026.0545.1
PLoPBase Model=LLaMA-3.1-8B, Params=125M2026.0543.3
Stochasticw_eff=322026.0426.8
SWAw_eff=1282026.0422.6
SWAw_eff=642026.043.7
SWAw_eff=322026.040