QASC Validation: Reranking Remains Stronger Than Either Hierarchical Or Adaptive Traversal (Results) in Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
This organizer groups the paper content from Results > Strict-Parity Retrieval > QASC validation: reranking remains stronger than either hierarchical or adaptive traversal. It contains three atomic results: (i) ColBERTv2/RePlug is the strongest strict-parity system at R@10 = 85.0 [83.4, 86.6]; (ii) Adaptive (heuristic) is statistically tied with the fixed-depth hierarchical baseline (ΔR@10 = +0.5 [−0.5, +1.5]); (iii) the paper therefore presents adaptive depth gating as an optional extension rather than the core result on QASC.
0
1
Tags
Science
Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
Related
LectureBank-Full: Diffusion Plus Quotas Account For The Main Gain; Adaptive Gating Is Optional (Results) in Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
MOOC-CS: Language Alignment Is The Bottleneck (Results) in Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
QASC Validation: Reranking Remains Stronger Than Either Hierarchical Or Adaptive Traversal (Results) in Auditable Strict-Parity Evaluation of Prerequisite-Graph Retrieval for RAG under Leakage Controls
Strict-Parity Retrieval Headline Takeaway: Three Durable Signals, Adaptive Depth Is Not The Main Story (Auditable Strict-Parity Graph-RAG Paper)