Prompt optimization on Penguins
68.6AccuracyTRAS
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| TRASLLM=GPT-3.5-turbo, Seeds=5, Iterations=15, Warmup steps (τwarmup)=3, K=62025.07 | 68.6 | 0.007 | 2.556 | |
| Baseline + TRLLM=GPT-3.5-turbo, Seeds=5, Iterations=15, Warmup steps (τwarmup)=32025.07 | 66.9 | 0.043 | 1.464 | |
| Baseline + MCSALLM=GPT-3.5-turbo, Seeds=5, Iterations=15, Warmup steps (τwarmup)=3, K=62025.07 | 66.1 | 0.083 | 1.148 | |
| BaselineLLM=GPT-3.5-turbo, Seeds=5, Iterations=152025.07 | 65.1 | — | — |