List your company

Question Answering on BBQ Overall Llama-3

80.7AccuracyCatRAG
47.62856.21464.873.386Mar 23, 2026

Evaluation results

MethodLinks
CatRAGModel=Llama-32026.0380.732.30.0197.6
Causal DebiasingModel=Llama-32026.0378.628.20.184.1
SP DebiasingModel=Llama-32026.0368.519.60.1969.2
CE DebiasingModel=Llama-32026.0364.215.30.2854.3
Self DebiasingModel=Llama-32026.0359.810.90.4135.1
Base ModelModel=Llama-32026.0348.900.630