ResearchBenchmarksVeracity Inference on PRONTOQA 4-hop (test)Follow96.7Hamming SimilarityAVI63.00471.75280.589.248May 17, 2025Evaluation ResultsMethodMethodLinksHamming SimilarityAVIBase LLM=Qwen 8BBase LLM=Qwen 8B2025.0596.7AVIBase LLM=Qwen 4BBase LLM=Qwen 4B2025.0592.1Many2ManyBase LLM=Qwen 4BBase LLM=Qwen 4B2025.0577.9Many2ManyBase LLM=Qwen 8BBase LLM=Qwen 8B2025.0564.3