Prompt Selection on Selected tasks APE-generated prompt pools APE design (test)
1.41Averaged Performance RankTRIPLE-CLST
Evaluation results
| Method | Links | |
|---|---|---|
| TRIPLE-CLSTcandidate_pool_size=150, total_budget=100, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 1.41 | |
| TRIPLE-CLSTcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 1.58 | |
| TRIPLE-GSEcandidate_pool_size=150, total_budget=100, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 1.58 | |
| TRIPLE-GSEcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 1.6 | |
| TRIPLE-CLSTcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 1.68 | |
| TRIPLE-GSEcandidate_pool_size=150, total_budget=100, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 1.75 | |
| TRIPLE-GSEcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 1.87 | |
| TRIPLE-SHcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=false2024.02 | 2 | |
| TRIPLE-CLSTcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=true2024.02 | 2 | |
| TRIPLE-SHcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=false2024.02 | 2.04 | |
| TRIPLE-SHcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=false2024.02 | 2.04 | |
| TRIPLE-CLSTcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 2.04 | |
| TRIPLE-CRcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=false2024.02 | 2.08 | |
| TRIPLE-CLSTcandidate_pool_size=150, total_budget=100, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 2.08 | |
| TRIPLE-GSEcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 2.16 | |
| TRIPLE-GSEcandidate_pool_size=150, total_budget=100, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 2.17 | |
| TRIPLE-CRcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=false2024.02 | 2.25 | |
| TRIPLE-CRcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=false2024.02 | 2.29 | |
| TRIPLE-CRcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=false2024.02 | 2.29 | |
| TRIPLE-CLSTcandidate_pool_size=150, total_budget=100, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 2.33 | |
| TRIPLE-SHcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=false2024.02 | 2.41 | |
| TRIPLE-GSEcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=true2024.02 | 2.41 | |
| UCBcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=false2024.02 | 2.46 | |
| UCBcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=false2024.02 | 2.5 | |
| UCBcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=false2024.02 | 2.5 | |
| NeuralUCBcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=true2024.02 | 2.58 | |
| BO-EIcandidate_pool_size=150, total_budget=100, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 2.58 | |
| UCBcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=false2024.02 | 2.66 | |
| BO-EIcandidate_pool_size=150, total_budget=100, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 2.75 | |
| BO-EIcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 2.91 | |
| BO-EIcandidate_pool_size=150, total_budget=100, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 2.91 | |
| NeuralUCBcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 3 | |
| BO-EIcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 3 | |
| Uniformcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=false2024.02 | 3 | |
| BO-EIcandidate_pool_size=30, total_budget=150, target_llm=Mistral, use_prompt_embeddings=true2024.02 | 3 | |
| Uniformcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=false2024.02 | 3.08 | |
| NeuralUCBcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 3.16 | |
| NeuralUCBcandidate_pool_size=150, total_budget=100, target_llm=Gemma, use_prompt_embeddings=true2024.02 | 3.16 | |
| Uniformcandidate_pool_size=30, total_budget=150, target_llm=Gemma, use_prompt_embeddings=false2024.02 | 3.21 | |
| NeuralUCBcandidate_pool_size=30, total_budget=150, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 3.25 | |
| Uniformcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=false2024.02 | 3.28 | |
| NeuralUCBcandidate_pool_size=150, total_budget=100, target_llm=Llama2, use_prompt_embeddings=true2024.02 | 3.5 | |
| BO-EIcandidate_pool_size=30, total_budget=150, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 3.67 | |
| NeuralUCBcandidate_pool_size=150, total_budget=100, target_llm=GPT 3.5, use_prompt_embeddings=true2024.02 | 3.75 |