# Multi-Hop Question Answering on 2Wiki Platinum (test) Source: [Multi-Hop Question Answering on 2Wiki Platinum (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:34:57.027Z Evaluation of model performance on the test split of the 2Wiki Platinum dataset for multi-hop question answering, using Ans and Unans metrics. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Multi-Hop Question Answering | | Dataset (including any stated split) | 2Wiki Platinum (test) | | Metric | Answer Rate | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 6 of 6 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Answer Rate](): Higher is better; unit: PERCENT. - [Unanswered Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Answer Rate | Result date | Report paper | | --- | --- | --- | --- | | PANINI | 84.8 | Not recorded | 699547c809c33900fcd6d721 | | HippoRAG 2 | 81.5 | Not recorded | 699547c809c33900fcd6d721 | | Qwen3 + reranker | 63.8 | Not recorded | 699547c809c33900fcd6d721 | | Qwen3 (8B) | 63.5 | Not recorded | 699547c809c33900fcd6d721 | | BM25 + reranker | 47.9 | Not recorded | 699547c809c33900fcd6d721 | | BM25 | 37.9 | Not recorded | 699547c809c33900fcd6d721 | ### PANINI Answer Rate: 84.8. Result ID: 699aa9c25698665ec4369cc4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Paradigm":"Structure-Augmented RAG","LLM Evaluator":"GPT-4o-mini"}. This record does not assert independent reproduction. ### HippoRAG 2 Answer Rate: 81.5. Result ID: 699aa9c25698665ec4369cc3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Paradigm":"Structure-Augmented RAG","LLM Evaluator":"GPT-4o-mini"}. This record does not assert independent reproduction. ### Qwen3 + reranker Answer Rate: 63.8. Result ID: 699aa9c25698665ec4369cc2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Paradigm":"Dense Retrieval","Backbone model":"Qwen3","Reranker usage":"True","LLM Evaluator":"GPT-4o-mini"}. This record does not assert independent reproduction. ### Qwen3 (8B) Answer Rate: 63.5. Result ID: 699aa9c25698665ec4369cc1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Paradigm":"Dense Retrieval","Backbone model":"Qwen3 (8B)","LLM Evaluator":"GPT-4o-mini"}. This record does not assert independent reproduction. ### BM25 + reranker Answer Rate: 47.9. Result ID: 699aa9c25698665ec4369cc0. [Reporting paper]() Recorded evaluation configuration: {"Retrieval Paradigm":"Sparse Retrieval","Reranker usage":"True","LLM Evaluator":"GPT-4o-mini"}. This record does not assert independent reproduction. ### BM25 Answer Rate: 37.9. Result ID: 699aa9c25698665ec4369cbf. [Reporting paper]() Recorded evaluation configuration: {"Retrieval Paradigm":"Sparse Retrieval","LLM Evaluator":"GPT-4o-mini"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.