List your company

Question Answering on CSQA (Accuracy)

88AccuracyMA-PoP
34.211248.175662.1476.1044Sep 26, 2025Nov 5, 2025Dec 15, 2025Jan 24, 2026Mar 5, 2026Apr 14, 2026May 25, 2026

Evaluation results

MethodLinks
MA-PoP2026.0588
LLP2026.0586.67
Sparse MADT=22026.0586.33
Self-Consistency2026.0586
ISP2026.0586
Free MADT=32026.0585.67
Decentr. MADT=32026.0585.33
Sparse MADT=32026.0585.33
Free MADT=22026.0585.33
Majority Voting2026.0585.33
Decentr. MADT=22026.0585
Sparse MADT=52026.0585
Free MADT=52026.0584.67
Decentr. MADT=52026.0584.33
Falcon-7B2026.0583
Centr. MADT=22026.0582.33
Centr. MADT=52026.0582.33
ICRBackbone=Qwen2.5-7B2025.0982
Qwen-7B2026.0581.53
ICVBackbone=Qwen2.5-7B2025.0981.3
IVBackbone=Qwen2.5-7B2025.0981.2
Centr. MADT=32026.0580.67
TVBackbone=Qwen2.5-7B2025.0980.6
Gemma-9B2026.0580.33
ELICITBackbone=Qwen2.5-7B2025.0979.2
FVBackbone=Qwen2.5-7B2025.0978.8
CommonSynBackbone=Llama-3.1-8B-Instruct, Fine-tuned on=CommonSyn, Evaluation Mode=Zero-shot2026.0370.8
VanillaBackbone=Llama-3.1-8B-Instruct, Evaluation Mode=Zero-shot2026.0370.3
BLT2026.0269.2
Mistral-7B2026.0568
Llama-8B2026.0567.67
OLMo 2Type=Sub-word reference2026.0266
OursTraining Stage=Stage 12026.0265.8
Bolmo2026.0264.2
OursTraining Stage=Stage 22026.0262.6
HnetTraining Stage=2-stage2026.0262.4
HnetTraining Stage=1-stage2026.0261.4
CommonGenBackbone=Llama-3.1-8B-Instruct, Fine-tuned on=CommonGen, Evaluation Mode=Zero-shot2026.0353.6
AdaMoLEFoundation Model=Qwen-2.5-3B, Jointly Fine-Tuning=true, r (per-expert rank)=6, n (number of experts)=4, k (number of activated experts)=22026.0549.39
RotMoLEFoundation Model=Qwen-2.5-3B, Jointly Fine-Tuning=true, r (per-expert rank)=6, n (number of experts)=4, k (number of activated experts)=22026.0543
MoLAFoundation Model=Qwen-2.5-3B, Jointly Fine-Tuning=true, r (per-expert rank)=6, n (number of experts)=4, k (number of activated experts)=22026.0539.31
MoCLEFoundation Model=Qwen-2.5-3B, Jointly Fine-Tuning=true, r (per-expert rank)=6, n (number of experts)=4, k (number of activated experts)=22026.0536.28