Safety and Truthfulness on Safety TruthfulQA CocoNot
68.97AccuracyParticipatory AI
Evaluation results
| Method | Links | |
|---|---|---|
| Participatory AImodularity=true, participation=true, number of models=322026.06 | 68.97 | |
| Participatory AImodularity=true, participation=true, number of models=162026.06 | 67.85 | |
| Participatory AImodularity=true, participation=true, number of models=82026.06 | 65.66 | |
| Participatory AImodularity=true, participation=true, number of models=42026.06 | 64.14 | |
| Participatory AImodularity=true, participation=true, number of models=22026.06 | 62.51 | |
| Modular, Non-participatory Baselinemodularity=true, participation=false2026.06 | 59.52 | |
| Non-modular, Non-participatory Baselinemodularity=false, participation=false2026.06 | 56.6 | |
| Non-modular, Participatory Baselinemodularity=false, participation=true2026.06 | 52.44 |