# Boosted Prompt Ensembles for Large Language Models

Source: [Boosted Prompt Ensembles for Large Language Models](<https://www.sota2.com/research/paper/68fa6f051e209131cea4352e>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-28T13:56:58.296Z

Authors: Silviu Pitis, Michael R. Zhang, Andrew Wang, Jimmy Ba

Published: 2023-04-12T16:47:15.000Z

## Abstract

Methods such as chain-of-thought prompting and self-consistency have pushed the frontier of language model reasoning performance with no additional training. To further improve performance, we propose a prompt ensembling method for large language models, which uses a small dataset to construct a set of few shot prompts that together comprise a \`\`boosted prompt ensemble''. The few shot examples for each prompt are chosen in a stepwise fashion to be \`\`hard'' examples on which the previous step's ensemble is uncertain. We show that this outperforms single-prompt output-space ensembles and bagged prompt-space ensembles on the GSM8k and AQuA datasets, among others. We propose both train-time and test-time versions of boosted prompting that use different levels of available annotation and conduct a detailed empirical study of our algorithm.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Task Progression Failure Detection on Cover Object Policy Success Rate: 3% (Out-of-Distribution)](<https://www.sota2.com/research/sota/task-progression-failure-detection-on-cover-object-policy-success-rate-3-percent-out-of-distribution.md>): Task Progression Failure Detection; Cover Object Policy Success Rate: 3% (Out-of-Distribution); TPR: 85.

- [Task progression failure detection on Cover Object (Combined)](<https://www.sota2.com/research/sota/task-progression-failure-detection-on-cover-object-combined.md>): Task Progression Failure Detection; Cover Object (Combined); TPR: 85.

- [Task Progression Failure Detection on Cover Object Policy Success Rate (In-Distribution)](<https://www.sota2.com/research/sota/task-progression-failure-detection-on-cover-object-policy-success-rate-in-distribution.md>): Task Progression Failure Detection; Cover Object Policy Success Rate: 98% (In-Distribution); TPR: 100.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2304.05970>)

[PDF](<https://arxiv.org/pdf/2304.05970.pdf>)