List your company

Closed-ended Task Evaluation on L-Eval closed-ended tasks

41.86Coursera ScoreResonance YaRN
28.714432.127235.5438.9528Feb 29, 2024

Evaluation results

MethodLinks
Resonance YaRNBase Model=LLaMA2-Chat 13B, Scaling Factor (s)=4, Fine-Tuning=FT@4K, 400 epochs, Context Length=16K2024.0241.863542.5765.85.5648.4439.87
YaRNBase Model=LLaMA2-Chat 13B, Scaling Factor (s)=4, Fine-Tuning=FT@4K, 400 epochs, Context Length=16K2024.0241.723441.0966.912.2248.4439.06
NTK-AwareBase Model=LLaMA2-Chat 13B, Scaling Factor (s)=4, Fine-Tuning=no FT, Context Length=16K2024.0240.262138.1265.431.1146.8835.47
Resonance YaRNBase Model=LLaMA2-Chat 13B, Scaling Factor (s)=4, Fine-Tuning=FT@16K, 100 epochs, Context Length=16K2024.0238.663943.5665.061.1162.541.65
YaRNBase Model=LLaMA2-Chat 13B, Scaling Factor (s)=4, Fine-Tuning=FT@16K, 100 epochs, Context Length=16K2024.0238.083943.0765.43063.2841.48
NTK-AwareBase Model=LLaMA2-Chat 7B, Scaling Factor (s)=8, Fine-Tuning=no FT, Context Length=32K2024.0236.77326.7334.21.1150.7825.43
Resonance YaRNBase Model=LLaMA2-Chat 7B, Scaling Factor (s)=8, Fine-Tuning=FT@32K, 50 epochs, Context Length=32K2024.0236.482234.1655.76057.0334.24
Resonance YaRNBase Model=LLaMA2-Chat 7B, Scaling Factor (s)=8, Fine-Tuning=FT@4K, 400 epochs, Context Length=32K2024.0236.342740.5956.513.3361.7237.58
YaRNBase Model=LLaMA2-Chat 7B, Scaling Factor (s)=8, Fine-Tuning=FT@32K, 50 epochs, Context Length=32K2024.0236.051933.1750.564.4456.2533.24
YaRNBase Model=LLaMA2-Chat 7B, Scaling Factor (s)=8, Fine-Tuning=FT@4K, 400 epochs, Context Length=32K2024.0235.032437.6257.624.4460.9436.61
Dynamic NTK-AwareBase Model=LLaMA2-Chat 7B, Fine-Tuning=no FT, Context Length=32K2024.0231.983234.6559.111.1136.7232.59
Dynamic NTK-AwareBase Model=LLaMA2-Chat 13B, Fine-Tuning=no FT, Context Length=16K2024.0229.223940.5963.941.1139.8435.62