Multi-agent Question Answering on PubMedQA (first 300 questions)
69.67Average AccuracyTextFullT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| TextFullTTraining Status=Training-Based, Communication Protocol=Text-Based, Number of Agents=3, Communication Turns=22026.05 | 69.67 | 71 | 505.03 | 5.47 | |
| HyLaTTraining Status=Training-Based, Communication Protocol=Latent-Space, Number of Agents=3, Communication Turns=22026.05 | 67.44 | 67.33 | 72.01 | 1.47 | |
| LatentFullTTraining Status=Training-Based, Communication Protocol=Latent-Space, Number of Agents=3, Communication Turns=22026.05 | 66.11 | 66.67 | 57 | 0.7 | |
| AutoFormTraining Status=Training-Free, Communication Protocol=Text-Based, Number of Agents=3, Communication Turns=22026.05 | 56.11 | 57 | 1,717.6 | 18.18 | |
| Single-AgentTraining Status=Training-Free, Communication Protocol=Single-Agent, Number of Agents=1, Communication Turns=02026.05 | 55.67 | 55.67 | 113.59 | 1.25 | |
| CipherTraining Status=Training-Free, Communication Protocol=Latent-Space, Number of Agents=3, Communication Turns=22026.05 | 55.44 | 57.67 | 1,137.89 | 11.58 | |
| EcoLangTraining Status=Training-Free, Communication Protocol=Text-Based, Number of Agents=3, Communication Turns=22026.05 | 54.67 | 57.33 | 960.26 | 10.04 | |
| NLTraining Status=Training-Free, Communication Protocol=Text-Based, Number of Agents=3, Communication Turns=22026.05 | 54.56 | 59 | 1,247.5 | 13.15 | |
| LatentMAS-VTraining Status=Training-Free, Communication Protocol=Latent-Space, Number of Agents=3, Communication Turns=22026.05 | 54.22 | 55 | 324.43 | 3.61 | |
| SDETraining Status=Training-Free, Communication Protocol=Text-Based, Number of Agents=3, Communication Turns=22026.05 | 49.67 | 53 | 1,316.2 | 14.48 |