List your company

Agent Tool Use on StableToolBench Held-In

50.4Pass RateSHAD+RFT
23.46430.45737.4544.443Dec 19, 2024

Evaluation results

MethodLinks
SHAD+RFTModel=LLaMA3.1-8B2024.1250.4
SHAD+RFTModel=LLaMA3-8B2024.1250.1
SHAD+α-FTModel=LLaMA3.1-8B2024.1249.2
SFTModel=LLaMA3.1-8B2024.1248.5
RewardFTModel=LLaMA3.1-8B2024.1248.2
SHAD+α-FTModel=LLaMA3-8B2024.1247
Regex+RFTModel=LLaMA3.1-8B2024.1246.7
RewardFTModel=LLaMA3-8B2024.1244.4
SFTModel=LLaMA3-8B2024.1243.1
RegexModel=LLaMA3.1-8B2024.1242.3
Regex+RFTModel=LLaMA3-8B2024.1241.2
RegexModel=LLaMA3-8B2024.1236.2
Rho-1Model=LLaMA3.1-8B2024.1230.6
Rho-1Model=LLaMA3-8B2024.1224.5