# Reactive Question Answering on StreamingBench excluding PO 1.0 (test) Source: [Reactive Question Answering on StreamingBench excluding PO 1.0 (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:54:31.773Z Evaluation of model performance on the specific test subset of StreamingBench, excluding PO 1.0, for Reactive Question Answering. 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 | Reactive Question Answering | | Dataset (including any stated split) | StreamingBench excluding PO 1.0 (test) | | Metric | Overall Performance (OP) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 9 of 9 | | 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 - [Overall Performance (OP)](): Higher is better; unit: PERCENT. - [Correct Responses (CR)](): Higher is better; unit: PERCENT. - [Contextual Score (CS)](): Higher is better; unit: PERCENT. - [Answerable Turn Precision (ATP)](): Higher is better; unit: PERCENT. - [Entity Understanding (EU)](): Higher is better; unit: PERCENT. - [Temporal Reasoning (TR)](): Higher is better; unit: PERCENT. - [Precision Rate (PR)](): Higher is better; unit: PERCENT. - [Subject Understanding (SU)](): Higher is better; unit: PERCENT. - [Answer Completeness (ACP)](): Higher is better; unit: PERCENT. - [Context Tracking (CT)](): Higher is better; unit: PERCENT. - [Error Rate (ER)](): Lower is better; unit: PERCENT. - [Stream Context Utilization (SCU)](): Higher is better; unit: PERCENT. - [Stream Dependency (SD)](): Higher is better; unit: PERCENT. - [Metric Accuracy (MA)](): Higher is better; unit: PERCENT. - [Answer Correctness Unit (ACU)](): Higher is better; unit: PERCENT. - [Multi-turn Context Utilization (MCU)](): Higher is better; unit: PERCENT. - [Streaming QA Score (SQA)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Performance (OP) | Result date | Report paper | | --- | --- | --- | --- | | ROMA | 76.96 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | ROMA | 76.69 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | ROMA | 76.13 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | ROMA | 75.61 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | ROMA | 75.61 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | ROMA | 75.51 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | Dispider | 74.92 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | VideoLLM-Online | 39.07 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | | Flash-VStream | 25.89 | Not recorded | 6969c5efe25ad2fa9d1ec7f7 | ### ROMA Overall Performance (OP): 76.96. Result ID: 6997b92e5698665ec40fa8b5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"wpos":3}. This record does not assert independent reproduction. ### ROMA Overall Performance (OP): 76.69. Result ID: 6997b92e5698665ec40fa8b8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"K":1}. This record does not assert independent reproduction. ### ROMA Overall Performance (OP): 76.13. Result ID: 6997b92e5698665ec40fa8b7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Ablation":"w/o Speak Head"}. This record does not assert independent reproduction. ### ROMA Overall Performance (OP): 75.61. Result ID: 6997b92e5698665ec40fa8b9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"wpos":2}. This record does not assert independent reproduction. ### ROMA Overall Performance (OP): 75.61. Result ID: 6997b92e5698665ec40fa8ba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"wpos":4}. This record does not assert independent reproduction. ### ROMA Overall Performance (OP): 75.51. Result ID: 6997b92e5698665ec40fa8b6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Ablation":"Mixed Training"}. This record does not assert independent reproduction. ### Dispider Overall Performance (OP): 74.92. Result ID: 6997b92e5698665ec40fa8b4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### VideoLLM-Online Overall Performance (OP): 39.07. Result ID: 6997b92e5698665ec40fa8b2. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Flash-VStream Overall Performance (OP): 25.89. Result ID: 6997b92e5698665ec40fa8b3. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.