List your company

Question Answering on TriviaQA 5shots

80Accuracyfp16
28.582441.931255.2868.6288Feb 7, 2026Feb 8, 2026Feb 9, 2026Feb 10, 2026Feb 11, 2026

Evaluation results

MethodLinks
fp16Size=70B2026.0280
MagRSize=70B2026.0277.5
AstroSize=70B2026.0277.5
OmniQuantSize=70B2026.0276.8
BNRMBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0271.99
fp16Size=13B2026.0270.4
FP16Backbone=LLaMA-3.1-8B, Quantization=None, Protocol=5-shot2026.0270.4
AstroSize=13B2026.0266.6
MagRSize=13B2026.0265.2
OmniQuantSize=13B2026.0264.5
fp16Size=7B2026.0264.4
AstroSize=7B2026.0255.9
MagRSize=7B2026.0255.3
OmniQuantSize=7B2026.0254.4
BNRMBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0254.3
SKBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0252.86
PoEBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0252.34
ALBMBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0252.09
ALBMBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0251.52
AstroBackbone=LLaMA-3.1-8B, Quantization=W3A16g128, Protocol=5-shot2026.0251.2
SKBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0249.01
BaseBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0248.53
PoEBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0248.41
LPBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0248.32
InfoRMBase Model=OpenRLHF-Llama3-8B-SFT, Evaluation Protocol=5-shot2026.0248.16
OmniQuantBackbone=LLaMA-3.1-8B, Quantization=W3A16g128, Protocol=5-shot2026.0247.6
LPBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0247.2
MagRBackbone=LLaMA-3.1-8B, Quantization=W3A16g128, Protocol=5-shot2026.0245.5
BaseBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0232.64
InfoRMBase Model=Llama3.1-8B-Instruct, Evaluation Protocol=5-shot2026.0230.56