# Multi-hop Question Answering on HybridQA (dev) Source: [Multi-hop Question Answering on HybridQA (dev)]() Collection index: [llms.txt]() Data updated: 2026-03-17T09:50:43.439Z Evaluation of model performance on the development split of the HybridQA dataset specifically for Multi-hop Question Answering, reporting Answer Correctness and Answer Completeness. 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) | HybridQA (dev) | | Metric | Answer Correctness | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 4 of 4 | | 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 Correctness](): Higher is better; unit: PERCENT. - [Answer Completeness](): Higher is better; unit: PERCENT. ## Recorded results | Method | Answer Correctness | Result date | Report paper | | --- | --- | --- | --- | | A.DOT | 71 | Not recorded | 69b8e03b0636e5bf4afbed40 | | Standard RAG | 56.2 | Not recorded | 69b8e03b0636e5bf4afbed40 | | ReAct | 40.2 | Not recorded | 69b8e03b0636e5bf4afbed40 | | LLM Compiler | 27.8 | Not recorded | 69b8e03b0636e5bf4afbed40 | ### A.DOT Answer Correctness: 71. Result ID: 69b9239b47f34ce924a23b8e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"LLaMA-3 70B","Evaluation Judge":"Mistral Large 2","Retrieval Architecture":"SQL + Milvus"}. This record does not assert independent reproduction. ### Standard RAG Answer Correctness: 56.2. Result ID: 69b9239b47f34ce924a23b8d. [Reporting paper]() Recorded evaluation configuration: {"Inference Model":"LLaMA-3 70B","Evaluation Judge":"Mistral Large 2","Retrieval Architecture":"SQL + Milvus"}. This record does not assert independent reproduction. ### ReAct Answer Correctness: 40.2. Result ID: 69b9239b47f34ce924a23b8c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Inference Model":"LLaMA-3 70B","Evaluation Judge":"Mistral Large 2","Retrieval Architecture":"SQL + Milvus"}. This record does not assert independent reproduction. ### LLM Compiler Answer Correctness: 27.8. Result ID: 69b9239b47f34ce924a23b8b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Inference Model":"LLaMA-3 70B","Evaluation Judge":"Mistral Large 2","Retrieval Architecture":"SQL + Milvus"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.