# Multi-document Multi-entity Question Answering on Loong 100K-200K Tokens Source: [Multi-document Multi-entity Question Answering on Loong 100K-200K Tokens]() Collection index: [llms.txt]() Data updated: 2026-03-13T08:53:20.495Z Evaluation of complex multi-document, multi-entity question answering capabilities on the Loong dataset variant constrained to 100K-200K token contexts, assessing performance across specific reasoning steps like Spotlight Locating, Comparison, Clustering, and Chain of Reason. 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-document Multi-entity Question Answering | | Dataset (including any stated split) | Loong 100K-200K Tokens | | Metric | Spotlight Locating Avg Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 5 of 5 | | 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 - [Spotlight Locating Avg Score](): Higher is better; unit: PERCENT. - [Spotlight Locating Perfect Rate](): Higher is better; unit: PERCENT. - [Comparison Avg Score](): Higher is better; unit: PERCENT. - [Comparison Perfect Rate](): Higher is better; unit: PERCENT. - [Clustering Avg Score](): Higher is better; unit: PERCENT. - [Clustering Perfect Rate](): Higher is better; unit: PERCENT. - [Chain of Reason Avg Score](): Higher is better; unit: PERCENT. - [Chain of Reason Perfect Rate](): Higher is better; unit: PERCENT. - [Overall Avg Score](): Higher is better; unit: PERCENT. - [Overall Perfect Rate](): Higher is better; unit: PERCENT. ## Recorded results | Method | Spotlight Locating Avg Score | Result date | Report paper | | --- | --- | --- | --- | | DocSage | 81.44 | Not recorded | 69b39a210636e5bf4afbcb2c | | GPT-4o | 76.34 | Not recorded | 69b39a210636e5bf4afbcb2c | | GPT-4o + RAG | 75.16 | Not recorded | 69b39a210636e5bf4afbcb2c | | StructRAG | 67.25 | Not recorded | 69b39a210636e5bf4afbcb2c | | GraphRAG | 16.15 | Not recorded | 69b39a210636e5bf4afbcb2c | ### DocSage Spotlight Locating Avg Score: 81.44. Result ID: 69b3d07047f34ce9249e066d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method":"DocSage (Ours)"}. This record does not assert independent reproduction. ### GPT-4o Spotlight Locating Avg Score: 76.34. Result ID: 69b3d07047f34ce9249e0669. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method":"GPT-4o"}. This record does not assert independent reproduction. ### GPT-4o + RAG Spotlight Locating Avg Score: 75.16. Result ID: 69b3d07047f34ce9249e066a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Method":"GPT-4o + RAG"}. This record does not assert independent reproduction. ### StructRAG Spotlight Locating Avg Score: 67.25. Result ID: 69b3d07047f34ce9249e066c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method":"StructRAG"}. This record does not assert independent reproduction. ### GraphRAG Spotlight Locating Avg Score: 16.15. Result ID: 69b3d07047f34ce9249e066b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Method":"GraphRAG"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.