List your company
ResearchFormerly WizWand

Hallucination Evaluation on TBH (ToolBench Hallucination)

61.6TBH ScoreRUBAS
-1.52814.86131.2547.639Jun 2, 2026
Updated 5mo ago

Evaluation results

MethodLinks
RUBASBase Model=Qwen3-8B, Training Strategy=+ RUBAS2026.0661.6
RUBASBase Model=Qwen3-14B, Training Strategy=+ RUBAS2026.0658.6
GuardModelBase Model=Qwen3-14B, Training Strategy=+ GuardModel2026.0657.3
RuleBase Model=Qwen3-14B, Training Strategy=+ Rule2026.0656.7
GuardModelBase Model=Qwen3-8B, Training Strategy=+ GuardModel2026.0656.3
RuleBase Model=Qwen3-8B, Training Strategy=+ Rule2026.0653.5
Qwen3-14BBase Model=Qwen3-14B, Training Strategy=Base2026.0642.1
Qwen3-8BBase Model=Qwen3-8B, Training Strategy=Base2026.0633.6
DPOBase Model=Qwen3-8B, Training Strategy=+ DPO2026.0629.8
DPOBase Model=Qwen3-14B, Training Strategy=+ DPO, Training Status=Failure (†)2026.0624.3
SFTBase Model=Qwen3-14B, Training Strategy=+ SFT2026.0623.2
RUBASBase Model=GLM-4.7-Flash, Training Strategy=+ RUBAS2026.0618.5
SFTBase Model=Qwen3-8B, Training Strategy=+ SFT2026.0617.4
SFTBase Model=GLM-4.7-Flash, Training Strategy=+ SFT2026.0610.5
RuleBase Model=GLM-4.7-Flash, Training Strategy=+ Rule, Training Status=Failure (†)2026.069.7
GLM-4.7-FlashBase Model=GLM-4.7-Flash, Training Strategy=Base2026.068.3
GuardModelBase Model=GLM-4.7-Flash, Training Strategy=+ GuardModel, Training Status=Failure (†)2026.066.4
DPOBase Model=GLM-4.7-Flash, Training Strategy=+ DPO, Training Status=Failure (†)2026.060.9