ResearchBenchmarksVeracity Inference on PRONTOQA 5-hop (test)Follow0.955Hamming SimilarityAVI0.65340.73170.810.8883May 17, 2025Evaluation ResultsMethodMethodLinksHamming SimilarityAVIBase LLM=Qwen 8BBase LLM=Qwen 8B2025.050.955AVIBase LLM=Qwen 4BBase LLM=Qwen 4B2025.050.913Many2ManyBase LLM=Qwen 4BBase LLM=Qwen 4B2025.050.684Many2ManyBase LLM=Qwen 8BBase LLM=Qwen 8B2025.050.665