List your company
ResearchFormerly WizWand

Commonsense on HellaSwag 10-shot

64.85Accuracy (10-shot)TuluDPO
55.635658.027860.4262.8122Nov 14, 2025
Updated 8mo ago

Evaluation results

MethodLinks
TuluDPOBase Model=Llama-3.1-8B-TuluSFT, Training Mixture/Protocol=TuluDPO2025.1164.85
UltraMix-190kBase Model=Llama-3.1-8B-TuluSFT, Training Mixture/Protocol=UM-190k2025.1164.82
UltraMix-187kBase Model=Llama-3.1-8B-TuluSFT, Training Mixture/Protocol=UM-187k2025.1164.02
UltraMix-190kBase Model=Qwen-2.5-7B-TuluSFT, Training Mixture/Protocol=UM-190k2025.1163.89
UltraMix-170kBase Model=Qwen-2.5-7B-TuluSFT, Training Mixture/Protocol=UM-170k2025.1163.59
UltraMix-187kBase Model=Qwen-2.5-7B-TuluSFT, Training Mixture/Protocol=UM-187k2025.1163.54
UltraMix-170kBase Model=Llama-3.1-8B-TuluSFT, Training Mixture/Protocol=UM-170k2025.1163.43
TuluDPOBase Model=Qwen-2.5-7B-TuluSFT, Training Mixture/Protocol=TuluDPO2025.1162.7
SFTBase Model=Qwen-2.5-7B-TuluSFT, Training Mixture/Protocol=SFT2025.1160.85
SFTBase Model=Llama-3.1-8B-TuluSFT, Training Mixture/Protocol=SFT2025.1160.41
Instella-3B-SFT (UM-190k)DPO Training Dataset=UM-190k, Evaluation Protocol=10-shot2025.1158.74
Instella-3B-SFT (UM-187k)DPO Training Dataset=UM-187k, Evaluation Protocol=10-shot2025.1158.56
Instella-3B-SFT (UM-170k)DPO Training Dataset=UM-170k, Evaluation Protocol=10-shot2025.1158.23
Instella-3B-SFT (TuluDPO)DPO Training Dataset=TuluDPO, Evaluation Protocol=10-shot2025.1157.51
Instella-3B-SFT (ORPO)DPO Training Dataset=ORPO, Evaluation Protocol=10-shot2025.1156.72
Instella-3B-SFT (UltraFB)DPO Training Dataset=UltraFB, Evaluation Protocol=10-shot2025.1156.3
Instella-3B-SFT (HelpSteer)DPO Training Dataset=HelpSteer, Evaluation Protocol=10-shot2025.1156.24
Instella-3B-SFT (SFT)DPO Training Dataset=None (SFT Baseline), Evaluation Protocol=10-shot2025.1156.2
Instella-3B-SFT (CodePref)DPO Training Dataset=CodePref, Evaluation Protocol=10-shot2025.1155.99