G2: Guided Generation for Enhanced Output Diversity in LLMs
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911245908574208 |
|---|---|
| author | Ruan, Zhiwen Li, Yixia Liu, Yefeng Chen, Yun Luo, Weihua Li, Peng Liu, Yang Chen, Guanhua |
| author_facet | Ruan, Zhiwen Li, Yixia Liu, Yefeng Chen, Yun Luo, Weihua Li, Peng Liu, Yang Chen, Guanhua |
| contents | Large Language Models (LLMs) have demonstrated exceptional performance across diverse natural language processing tasks. However, these models exhibit a critical limitation in output diversity, often generating highly similar content across multiple attempts. This limitation significantly affects tasks requiring diverse outputs, from creative writing to reasoning. Existing solutions, like temperature scaling, enhance diversity by modifying probability distributions but compromise output quality. We propose Guide-to-Generation (G2), a training-free plug-and-play method that enhances output diversity while preserving generation quality. G2 employs a base generator alongside dual Guides, which guide the generation process through decoding-based interventions to encourage more diverse outputs conditioned on the original query. Comprehensive experiments demonstrate that G2 effectively improves output diversity while maintaining an optimal balance between diversity and quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_00432 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | G2: Guided Generation for Enhanced Output Diversity in LLMs Ruan, Zhiwen Li, Yixia Liu, Yefeng Chen, Yun Luo, Weihua Li, Peng Liu, Yang Chen, Guanhua Computation and Language Large Language Models (LLMs) have demonstrated exceptional performance across diverse natural language processing tasks. However, these models exhibit a critical limitation in output diversity, often generating highly similar content across multiple attempts. This limitation significantly affects tasks requiring diverse outputs, from creative writing to reasoning. Existing solutions, like temperature scaling, enhance diversity by modifying probability distributions but compromise output quality. We propose Guide-to-Generation (G2), a training-free plug-and-play method that enhances output diversity while preserving generation quality. G2 employs a base generator alongside dual Guides, which guide the generation process through decoding-based interventions to encourage more diverse outputs conditioned on the original query. Comprehensive experiments demonstrate that G2 effectively improves output diversity while maintaining an optimal balance between diversity and quality. |
| title | G2: Guided Generation for Enhanced Output Diversity in LLMs |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2511.00432 |