List your company

Agent Tool Use on Nexus (Held-Out)

32AccuracySHAD+RFT
5.42812.326519.22526.1235Dec 19, 2024

Evaluation results

MethodLinks
SHAD+RFTModel=LLaMA3.1-8B2024.1232
SHAD+α-FTModel=LLaMA3.1-8B2024.1228.9
SHAD+α-FTModel=LLaMA3-8B2024.1228.7
SHAD+RFTModel=LLaMA3-8B2024.1227.8
Rho-1Model=LLaMA3.1-8B2024.1226
SFTModel=LLaMA3.1-8B2024.1219.5
RewardFTModel=LLaMA3.1-8B2024.1219.1
Rho-1Model=LLaMA3-8B2024.1219
Regex+RFTModel=LLaMA3.1-8B2024.1216.2
RegexModel=LLaMA3.1-8B2024.1214.3
SFTModel=LLaMA3-8B2024.1214
Regex+RFTModel=LLaMA3-8B2024.1212.4
RewardFTModel=LLaMA3-8B2024.128
RegexModel=LLaMA3-8B2024.126.45