Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2410.07737 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910643848740864 |
|---|---|
| author | Wang, Can Sui, Dianbo Sun, Hongliang Ding, Hao Zhang, Bolin Tu, Zhiying |
| author_facet | Wang, Can Sui, Dianbo Sun, Hongliang Ding, Hao Zhang, Bolin Tu, Zhiying |
| contents | Large Language Model (LLM) services exhibit impressive capability on unlearned tasks leveraging only a few examples by in-context learning (ICL). However, the success of ICL varies depending on the task and context, leading to heterogeneous service quality. Directly estimating the performance of LLM services at each invocation can be laborious, especially requiring abundant labeled data or internal information within the LLM. This paper introduces a novel method to estimate the performance of LLM services across different tasks and contexts, which can be "plug-and-play" utilizing only a few unlabeled samples like ICL. Our findings suggest that the negative log-likelihood and perplexity derived from LLM service invocation can function as effective and significant features. Based on these features, we utilize four distinct meta-models to estimate the performance of LLM services. Our proposed method is compared against unlabeled estimation baselines across multiple LLM services and tasks. And it is experimentally applied to two scenarios, demonstrating its effectiveness in the selection and further optimization of LLM services. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_07737 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Plug-and-Play Performance Estimation for LLM Services without Relying on Labeled Data Wang, Can Sui, Dianbo Sun, Hongliang Ding, Hao Zhang, Bolin Tu, Zhiying Performance Machine Learning Large Language Model (LLM) services exhibit impressive capability on unlearned tasks leveraging only a few examples by in-context learning (ICL). However, the success of ICL varies depending on the task and context, leading to heterogeneous service quality. Directly estimating the performance of LLM services at each invocation can be laborious, especially requiring abundant labeled data or internal information within the LLM. This paper introduces a novel method to estimate the performance of LLM services across different tasks and contexts, which can be "plug-and-play" utilizing only a few unlabeled samples like ICL. Our findings suggest that the negative log-likelihood and perplexity derived from LLM service invocation can function as effective and significant features. Based on these features, we utilize four distinct meta-models to estimate the performance of LLM services. Our proposed method is compared against unlabeled estimation baselines across multiple LLM services and tasks. And it is experimentally applied to two scenarios, demonstrating its effectiveness in the selection and further optimization of LLM services. |
| title | Plug-and-Play Performance Estimation for LLM Services without Relying on Labeled Data |
| topic | Performance Machine Learning |
| url | https://arxiv.org/abs/2410.07737 |