# Multi-agent Question Answering on SocialIQA (first 300 questions) Source: [Multi-agent Question Answering on SocialIQA (first 300 questions)]() Collection index: [llms.txt]() Data updated: 2026-05-26T10:59:10.076Z Evaluation of model performance on the first 300 questions subset of the SocialIQA dataset for Multi-agent Question Answering, including accuracy and efficiency 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-agent Question Answering | | Dataset (including any stated split) | SocialIQA (first 300 questions) | | Metric | Average Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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 - [Average Accuracy](): Higher is better; unit: PERCENT. - [Majority Accuracy](): Higher is better; unit: PERCENT. - [Token Count](): Higher is better; unit: SCALAR. - [Inference Time](): Lower is better; unit: TIME. ## Recorded results | Method | Average Accuracy | Result date | Report paper | | --- | --- | --- | --- | | HyLaT | 82.56 | Not recorded | 6a152968dbd7150506f7e36a | | LatentFullT | 80.78 | Not recorded | 6a152968dbd7150506f7e36a | | TextFullT | 76.66 | Not recorded | 6a152968dbd7150506f7e36a | | AutoForm | 59.89 | Not recorded | 6a152968dbd7150506f7e36a | | NL | 58.22 | Not recorded | 6a152968dbd7150506f7e36a | | SDE | 58.11 | Not recorded | 6a152968dbd7150506f7e36a | | Cipher | 57.89 | Not recorded | 6a152968dbd7150506f7e36a | | Single-Agent | 56 | Not recorded | 6a152968dbd7150506f7e36a | | EcoLang | 55.78 | Not recorded | 6a152968dbd7150506f7e36a | | LatentMAS-V | 42.78 | Not recorded | 6a152968dbd7150506f7e36a | ### HyLaT Average Accuracy: 82.56. Result ID: 6a157cccfbc373592c0dbca4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Based","Communication Protocol":"Latent-Space","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### LatentFullT Average Accuracy: 80.78. Result ID: 6a157cccfbc373592c0dbca3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Based","Communication Protocol":"Latent-Space","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### TextFullT Average Accuracy: 76.66. Result ID: 6a157cccfbc373592c0dbca2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Based","Communication Protocol":"Text-Based","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### AutoForm Average Accuracy: 59.89. Result ID: 6a157cccfbc373592c0dbc9d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Text-Based","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### NL Average Accuracy: 58.22. Result ID: 6a157cccfbc373592c0dbc9c. [Reporting paper]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Text-Based","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### SDE Average Accuracy: 58.11. Result ID: 6a157cccfbc373592c0dbc9f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Text-Based","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### Cipher Average Accuracy: 57.89. Result ID: 6a157cccfbc373592c0dbca0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Latent-Space","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### Single-Agent Average Accuracy: 56. Result ID: 6a157cccfbc373592c0dbc9b. [Reporting paper]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Single-Agent","Number of Agents":"1","Communication Turns":"0"}. This record does not assert independent reproduction. ### EcoLang Average Accuracy: 55.78. Result ID: 6a157cccfbc373592c0dbc9e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Text-Based","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. ### LatentMAS-V Average Accuracy: 42.78. Result ID: 6a157cccfbc373592c0dbca1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Training Status":"Training-Free","Communication Protocol":"Latent-Space","Number of Agents":"3","Communication Turns":"2"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.