Why is prompting hard? Understanding prompts on binary sequence predictors
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914552813191168 |
|---|---|
| author | Wenliang, Li Kevin Ruoss, Anian Grau-Moya, Jordi Hutter, Marcus Genewein, Tim |
| author_facet | Wenliang, Li Kevin Ruoss, Anian Grau-Moya, Jordi Hutter, Marcus Genewein, Tim |
| contents | Frontier models can be prompted or conditioned to do many tasks, but finding good prompts is not always easy, nor is understanding some performant prompts. We view prompting as finding the best conditioning sequence on a near-optimal sequence predictor. On numerous well-controlled experiments, we show that unintuitive optimal conditioning sequences can be better understood given the pretraining distribution, which is not usually available. Even using exhaustive search, reliably identifying optimal prompts for practical neural predictors can be surprisingly difficult. Popular prompting methods, such as using demonstrations from the targeted task, can be surprisingly suboptimal. Using the same empirical framework, we analyze optimal prompts on frontier models, revealing patterns similar to the binary examples and previous findings. Taken together, this work takes an initial step towards understanding optimal prompts, from a statistical and empirical perspective that complements research on frontier models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_10760 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Why is prompting hard? Understanding prompts on binary sequence predictors Wenliang, Li Kevin Ruoss, Anian Grau-Moya, Jordi Hutter, Marcus Genewein, Tim Computation and Language Machine Learning Frontier models can be prompted or conditioned to do many tasks, but finding good prompts is not always easy, nor is understanding some performant prompts. We view prompting as finding the best conditioning sequence on a near-optimal sequence predictor. On numerous well-controlled experiments, we show that unintuitive optimal conditioning sequences can be better understood given the pretraining distribution, which is not usually available. Even using exhaustive search, reliably identifying optimal prompts for practical neural predictors can be surprisingly difficult. Popular prompting methods, such as using demonstrations from the targeted task, can be surprisingly suboptimal. Using the same empirical framework, we analyze optimal prompts on frontier models, revealing patterns similar to the binary examples and previous findings. Taken together, this work takes an initial step towards understanding optimal prompts, from a statistical and empirical perspective that complements research on frontier models. |
| title | Why is prompting hard? Understanding prompts on binary sequence predictors |
| topic | Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2502.10760 |