Through reading the documentation in the context, tool-using language models can dynamically extend their capability using external tools. The cost is that we have to input lengthy documentation every time the model needs to use the tool, occupying the input window as well as slowing down the decoding process.
Given the progress in general-purpose compression, soft context compression is a suitable approach to alleviate the problem. However, when compressing tool documentation, existing methods suffer from the weaknesses of key information loss (specifically, tool/parameter name errors) and difficulty in adjusting the length of compressed sequences based on documentation lengths.
To address these problems, we propose two strategies for compressing tool documentation into concise and precise summary sequences for tool-using language models. 1) Selective compression strategy mitigates key information loss by deliberately retaining key information as raw text tokens. 2) Block compression strategy involves dividing tool documentation into short chunks and then employing a fixed-length compression model to achieve variable-length compression. This strategy facilitates the flexible adjustment of the compression ratio.
Results on API-Bank and APIBench show that our approach reaches a performance comparable to the upper-bound baseline under up to 16x compression ratio.
Efficient Learned Data Compression via Dual-Stream Feature Decoupling
College of Computer Science, TMCC, SysNet, DISSec, GTIISC, Nankai University(南开大学计算机学院,TMCC,SysNet,DISSec,GTIISC);College of Computing and Data Science, Nanyang Technological University(南洋理工大学计算与数据科学学院);Beijing Institute of Technology(北京理工大学)
Adaptive Feature-based Low-Rank Compression of Large Language Models via Bayesian Optimization
School of Computer Science and Technology, Soochow University(苏州大学计算机科学与技术学院);Huawei Cloud, China(华为云);Harbin Institute of Technology, Shenzhen(哈尔滨工业大学(深圳))