List your company

Question Answering on HaluEval QA

45.4AccuracyHaluSearch (MCTS)
27.51232.15636.841.444Jan 2, 2025

Evaluation results

MethodLinks
HaluSearch (MCTS)Backbone=Llama3.1-8B-Instruct2025.0145.4
HaluSearch (MCTS)Backbone=Qwen2-7B-Instruct2025.0143.69
BoNBackbone=Qwen2-7B-Instruct2025.0139
BoNBackbone=Llama3.1-8B-Instruct2025.0137.8
SCBackbone=Qwen2-7B-Instruct2025.0137.2
CoTBackbone=Llama3.1-8B-Instruct2025.0136.8
Self-RefineBackbone=Qwen2-7B-Instruct2025.0135.8
Direct GenerationBackbone=Llama3.1-8B-Instruct2025.0135.6
CoTBackbone=Qwen2-7B-Instruct2025.0135.6
ITIBackbone=Llama3.1-8B-Instruct2025.0135.2
SCBackbone=Llama3.1-8B-Instruct2025.0135
Direct GenerationBackbone=Qwen2-7B-Instruct2025.0134
ITIBackbone=Qwen2-7B-Instruct2025.0132.2
Self-RefineBackbone=Llama3.1-8B-Instruct2025.0128.2