chain-of-draft
If you are the rightful owner of chain-of-draft and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
The Chain of Draft (CoD) MCP Server implements a novel reasoning approach for LLMs, optimizing token usage while maintaining accuracy.
The Chain of Draft (CoD) MCP Server is designed to implement the Chain of Draft reasoning approach, which is a novel paradigm that allows language models to generate minimalistic yet informative intermediate reasoning outputs. This approach significantly reduces token usage while maintaining accuracy, making it efficient and cost-effective. CoD is applicable across various reasoning tasks and domains, offering flexibility and adaptability. The server supports both Python and JavaScript implementations, ensuring compatibility with different platforms and easy integration into existing workflows.
Features
- Core Chain of Draft Implementation: Concise reasoning steps with format enforcement and answer extraction.
- Performance Analytics: Tracks token usage, solution accuracy, execution time, and domain-specific performance metrics.
- Adaptive Word Limits: Automatically estimates complexity and adjusts word limits dynamically.
- Comprehensive Example Database: Transforms CoT examples to CoD format with domain-specific examples.
- Hybrid Reasoning Approaches: Automatically selects between CoD and CoT based on domain-specific optimization.
Tools
chain_of_draft_solve
Solve a problem using Chain of Draft reasoning.
math_solve
Solve a math problem with CoD.
code_solve
Solve a coding problem with CoD.
logic_solve
Solve a logic problem with CoD.
get_performance_stats
Get performance stats for CoD vs CoT.
