Model Discovery on Llama-8B model tree Extended Discovery
1RankOptimal
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OptimalQueries Per Model=N/A2026.01 | 1 | 74.7 | |
| accelerated Sequential HalvingQueries Per Model=2002026.01 | 1.3 | 74.4 | |
| UCB-StdDev.Queries Per Model=2002026.01 | 1.8 | 74.3 | |
| accelerated Sequential HalvingQueries Per Model=1002026.01 | 1.9 | 74 | |
| UCB-EQueries Per Model=2002026.01 | 1.9 | 74.3 | |
| accelerated Sequential HalvingQueries Per Model=502026.01 | 3 | 73.6 | |
| BayesElimQueries Per Model=2002026.01 | 4 | 73.8 | |
| Sequential HalvingQueries Per Model=2002026.01 | 4.3 | 73.9 | |
| accelerated Sequential HalvingQueries Per Model=252026.01 | 8 | 73 | |
| UCBQueries Per Model=2002026.01 | 8.4 | 74.1 | |
| BayesElimQueries Per Model=1002026.01 | 11.6 | 72.9 | |
| UCB-StdDev.Queries Per Model=1002026.01 | 13.2 | 73.4 | |
| UCB-EQueries Per Model=1002026.01 | 13.2 | 73.5 | |
| Sequential HalvingQueries Per Model=1002026.01 | 14.8 | 72.6 | |
| accelerated Sequential HalvingQueries Per Model=102026.01 | 15.4 | 72.5 | |
| TTTSQueries Per Model=2002026.01 | 16.1 | 73.4 | |
| UCBQueries Per Model=1002026.01 | 27.7 | 72.8 | |
| Sequential HalvingQueries Per Model=502026.01 | 29.9 | 72 | |
| BayesElimQueries Per Model=502026.01 | 34.4 | 72.1 | |
| UCB-StdDev.Queries Per Model=502026.01 | 37.1 | 72.2 | |
| Best BaseQueries Per Model=N/A2026.01 | 41 | 71.3 | |
| UCB-EQueries Per Model=502026.01 | 43.4 | 72 | |
| TTTSQueries Per Model=1002026.01 | 46.6 | 71.8 | |
| Sequential HalvingQueries Per Model=252026.01 | 50.9 | 71.7 | |
| Successive RejectsQueries Per Model=2002026.01 | 52.2 | 71.6 | |
| UniformQueries Per Model=2002026.01 | 54.8 | 72.3 | |
| UCB-StdDev.Queries Per Model=252026.01 | 55.4 | 71.8 | |
| UCBQueries Per Model=502026.01 | 55.8 | 71.8 | |
| BayesElimQueries Per Model=252026.01 | 56.3 | 71.5 | |
| Successive RejectsQueries Per Model=1002026.01 | 63.3 | 71.3 | |
| UCB-EQueries Per Model=252026.01 | 63.7 | 71.5 | |
| TTTSQueries Per Model=502026.01 | 70.6 | 71.2 | |
| UCBQueries Per Model=252026.01 | 74.1 | 71.2 | |
| Sequential HalvingQueries Per Model=102026.01 | 75.2 | 71 | |
| BayesElimQueries Per Model=102026.01 | 79.9 | 70.8 | |
| UniformQueries Per Model=1002026.01 | 86.2 | 71.3 | |
| Successive RejectsQueries Per Model=502026.01 | 92 | 70.7 | |
| UCB-StdDev.Queries Per Model=102026.01 | 96 | 70.7 | |
| TTTSQueries Per Model=252026.01 | 96.2 | 70.5 | |
| TTTSQueries Per Model=102026.01 | 103.6 | 70.2 | |
| UCB-EQueries Per Model=102026.01 | 106.5 | 70.3 | |
| UCBQueries Per Model=102026.01 | 110.2 | 70.2 | |
| UniformQueries Per Model=502026.01 | 114.9 | 70.5 | |
| Successive RejectsQueries Per Model=252026.01 | 122.3 | 69.9 | |
| UniformQueries Per Model=252026.01 | 167.9 | 68.7 | |
| Successive RejectsQueries Per Model=102026.01 | 171.5 | 68.3 | |
| UniformQueries Per Model=102026.01 | 222 | 66.5 | |
| Random SelectionQueries Per Model=N/A2026.01 | 270.8 | 60.1 |