Efficient Multi-objective Prompt Optimization via Pure-exploration Bandits
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914565855379456 |
|---|---|
| author | Li, Donghao Shi, Chengshuai Ou, Weijuan Shen, Cong Yang, Jing |
| author_facet | Li, Donghao Shi, Chengshuai Ou, Weijuan Shen, Cong Yang, Jing |
| contents | Prompt engineering has become central to eliciting the capabilities of large language models (LLMs). At its core lies prompt selection -- efficiently identifying the most effective prompts. However, most prior investigations overlook a key challenge: the inherently multi-faceted nature of prompt performance, which cannot be captured by a single metric. To fill this gap, we study the multi-objective prompt selection problem under two practical settings: Pareto prompt set recovery and best feasible prompt identification. Casting the problem into the pure-exploration bandits framework, we adapt provably efficient algorithms from multi-objective bandits and further introduce a novel design for best feasible arm identification in structured bandits, with theoretical guarantees on the identification error in the linear case. Extensive experiments across multiple LLMs show that the bandit-based approaches yield significant improvements over baselines, establishing a principled and efficient framework for multi-objective prompt optimization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_14553 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Efficient Multi-objective Prompt Optimization via Pure-exploration Bandits Li, Donghao Shi, Chengshuai Ou, Weijuan Shen, Cong Yang, Jing Machine Learning Artificial Intelligence Prompt engineering has become central to eliciting the capabilities of large language models (LLMs). At its core lies prompt selection -- efficiently identifying the most effective prompts. However, most prior investigations overlook a key challenge: the inherently multi-faceted nature of prompt performance, which cannot be captured by a single metric. To fill this gap, we study the multi-objective prompt selection problem under two practical settings: Pareto prompt set recovery and best feasible prompt identification. Casting the problem into the pure-exploration bandits framework, we adapt provably efficient algorithms from multi-objective bandits and further introduce a novel design for best feasible arm identification in structured bandits, with theoretical guarantees on the identification error in the linear case. Extensive experiments across multiple LLMs show that the bandit-based approaches yield significant improvements over baselines, establishing a principled and efficient framework for multi-objective prompt optimization. |
| title | Efficient Multi-objective Prompt Optimization via Pure-exploration Bandits |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2605.14553 |