Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2407.19405 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917069855916032 |
|---|---|
| author | Chen, Dong Zhang, Shilin Gao, Fei Zhuang, Yueting Tang, Siliang Liu, Qidong Xu, Mingliang |
| author_facet | Chen, Dong Zhang, Shilin Gao, Fei Zhuang, Yueting Tang, Siliang Liu, Qidong Xu, Mingliang |
| contents | Large language models (LLMs) have garnered increasing attention owing to their powerful logical reasoning capabilities. Generally, larger LLMs (L-LLMs) that require paid interfaces exhibit significantly superior performance compared to smaller LLMs (S-LLMs) that can be deployed on a variety of devices. Knowledge distillation (KD) aims to empower S-LLMs with the capabilities of L-LLMs, while S-LLMs merely mimic the outputs of L-LLMs, failing to get the powerful logical reasoning capabilities. Consequently, S-LLMs are helpless when it comes to planning and decision-making tasks that require logical reasoning capabilities. To tackle the identified challenges, we propose a novel framework called Logic Distillation (LD). Initially, LD employs L-LLMs to instantiate complex instructions into discrete functions and illustrates their usage to establish a function base. Subsequently, based on the function base, LD fine-tunes S-LLMs to learn the logic employed by L-LLMs in planning and decision-making. During testing, LD utilizes a retriever to identify the top-$K$ relevant functions based on instructions and current states, which will be selected and invoked by S-LLMs. Ultimately, S-LLMs yield planning and decision-making outcomes, function by function. Relevant experiments demonstrate that with the assistance of LD, S-LLMs can achieve outstanding results in planning and decision-making tasks, comparable to, or even surpassing, those of L-LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_19405 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Logic Distillation: Learning from Code Function by Function for Decision-making Tasks Chen, Dong Zhang, Shilin Gao, Fei Zhuang, Yueting Tang, Siliang Liu, Qidong Xu, Mingliang Artificial Intelligence Large language models (LLMs) have garnered increasing attention owing to their powerful logical reasoning capabilities. Generally, larger LLMs (L-LLMs) that require paid interfaces exhibit significantly superior performance compared to smaller LLMs (S-LLMs) that can be deployed on a variety of devices. Knowledge distillation (KD) aims to empower S-LLMs with the capabilities of L-LLMs, while S-LLMs merely mimic the outputs of L-LLMs, failing to get the powerful logical reasoning capabilities. Consequently, S-LLMs are helpless when it comes to planning and decision-making tasks that require logical reasoning capabilities. To tackle the identified challenges, we propose a novel framework called Logic Distillation (LD). Initially, LD employs L-LLMs to instantiate complex instructions into discrete functions and illustrates their usage to establish a function base. Subsequently, based on the function base, LD fine-tunes S-LLMs to learn the logic employed by L-LLMs in planning and decision-making. During testing, LD utilizes a retriever to identify the top-$K$ relevant functions based on instructions and current states, which will be selected and invoked by S-LLMs. Ultimately, S-LLMs yield planning and decision-making outcomes, function by function. Relevant experiments demonstrate that with the assistance of LD, S-LLMs can achieve outstanding results in planning and decision-making tasks, comparable to, or even surpassing, those of L-LLMs. |
| title | Logic Distillation: Learning from Code Function by Function for Decision-making Tasks |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2407.19405 |