ResearchBenchmarksMulti-hop Question Answering on InfoQA Synthetic Multi-hop QA (4-hop)Follow80Avg F1InfoQA-3.218.44061.6Sep 25, 2025Evaluation ResultsMethodMethodLinksAvg F1InfoQAContext Length=Average...Context Length=Average, Backbone=Qwen3-14B2025.0980InfoQAContext Length=10k, Ba...Context Length=10k, Backbone=Qwen3-14B2025.0930DirectContext Length=Average...Context Length=Average, Backbone=Qwen3-14B2025.099DirectContext Length=10k, Ba...Context Length=10k, Backbone=Qwen3-14B2025.090