List your company

Commonsense Reasoning on HellaSwag (Score)

95.45HellaSwag ScoreLD-MoLE
28.245245.692663.1480.5874Feb 16, 2025Apr 23, 2025Jun 28, 2025Sep 2, 2025Nov 7, 2025Jan 12, 2026Mar 19, 2026

Evaluation results

MethodLinks
LD-MoLEModel=Llama3.1-8B, TP=2.33 %2025.0995.45
ReMoLEModel=Llama3.1-8B, TP=2.17 %2025.0994.95
MoLA(8888)Model=Llama3.1-8B, TP=2.17 %2025.0994.06
MoLA(2468)Model=Llama3.1-8B, TP=1.26 %2025.0994.03
LD-MoLEModel=Llama3.2-3B, TP=3.28 %2025.0993.6
ReMoLEModel=Llama3.2-3B, TP=3.11 %2025.0993.44
Trinity Large Baseshot=5-shot2026.0290.11
LD-MoLEModel=Qwen3-1.7B, TP=4.23 %2025.0988.71
ReMoLEModel=Qwen3-1.7B, TP=4.12 %2025.0988
MoLA(8888)Model=Llama3.2-3B, TP=3.11 %2025.0987.47
MoLA(2468)Model=Llama3.2-3B, TP=1.80 %2025.0987.31
MoLA(2468)Model=Qwen3-1.7B, TP=2.39 %2025.0987.09
MoLA(8888)Model=Qwen3-1.7B, TP=4.12 %2025.0983.35
FSWModel=Mistral-7B-v0.32026.0369
Clean FTModel=LLaMA-3-8B2026.0368.6
Clean FTModel=Mistral-7B-v0.32026.0368.6
FSWModel=LLaMA-3-8B2026.0367.8
Clean FTModel=DeepSeek-7B-Chat2026.0367.4
Clean FTModel=Qwen2.5-7B2026.0367
FSWModel=Qwen2.5-7B2026.0366.8
Clean FTModel=LLaMA-2-7B-hf2026.0366.4
FSWModel=LLaMA-2-7B-hf2026.0364.8
UM-190kBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1161.62
UM-187kBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1160.69
FSWModel=DeepSeek-7B-Chat2026.0360.4
UM-170kBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1159.9
TuluDPOBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1159.52
AdaGCBackbone=Llama-2 7B, Zero-Shot=true2025.0257.94
ORPOBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1157.82
SFTBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1157.32
UltraFBBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1157.14
HelpSteerBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1156.9
GlobalGCBackbone=Llama-2 7B, Zero-Shot=true2025.0256.06
AdaGCBackbone=Mixtral 8x1B, Zero-Shot=true2025.0255.85
AGCBackbone=Llama-2 7B, Zero-Shot=true2025.0255.47
UM-190kshots=10-shot2025.1155.09
UM-187kshots=10-shot2025.1154.93
UltraFBshots=10-shot2025.1153.98
ORPOshots=10-shot2025.1153.69
ClipByValueBackbone=Llama-2 7B, Zero-Shot=true2025.0253.49
TuluDPOshots=10-shot2025.1153.42
ClippyBackbone=Llama-2 7B, Zero-Shot=true2025.0253.34
GlobalGCBackbone=Mixtral 8x1B, Zero-Shot=true2025.0253.08
SFTshots=10-shot2025.1152.91
UM-170kshots=10-shot2025.1152.84
HelpSteershots=10-shot2025.1152.73
CodePrefBackbone=Apertus-8B-SFT, Evaluation Protocol=10-shot2025.1151.31
CodePrefshots=10-shot2025.1150.23
AdaGCBackbone=Llama-2 1.3B, Zero-Shot=true2025.0247.29
GlobalGCBackbone=Llama-2 1.3B, Zero-Shot=true2025.0246.62
ClippyBackbone=Llama-2 1.3B, Zero-Shot=true2025.0245.43
ClipByValueBackbone=Llama-2 1.3B, Zero-Shot=true2025.0244.11
BYOL-mriParameters=12B, Training Stage=CPT2026.0138.11
BYOL-mriParameters=4B, Training Stage=CPT2026.0137.8
BYOL-mriParameters=1B, Training Stage=CPT2026.0130.83