List your company

Hallucination Prediction on Overshadowing

65Accuracy (Time)CoDA
-1.97615.41232.850.188Feb 22, 2025

Evaluation results

MethodLinks
CoDABackbone=Llama-2-7b-chat2025.026546.8—
CoDABackbone=Mistral-7b2025.0261.249.5—
DolaBackbone=Llama-2-7b-chat2025.0253.631.8—
DolaBackbone=Mistral-7b2025.025134.6—
AlarmerModel=Mistral 7b, Protocol=Alarmer2025.0242.551.6—
SRBackbone=Llama-2-7b-chat2025.0242.523.8—
SRBackbone=Mistral-7b2025.0242.424.9—
GreedyBackbone=Llama-2-7b-chat2025.0241.420.8—
AlarmerModel=Llama-2-7b-chat, Protocol=Alarmer2025.0240.453.3—
CoTBackbone=Llama-2-7b-chat2025.0240.4——
USCBackbone=Llama-2-7b-chat2025.0240.216.4—
GreedyBackbone=Mistral-7b2025.0239.521.6—
USCBackbone=Mistral-7b2025.0237.920.8—
CoTBackbone=Mistral-7b2025.0237——
PromptModel=Mistral 7b, Protocol=Prompt2025.022.2——
PromptModel=Llama-2-7b-chat, Protocol=Prompt2025.020.6——