Skip to content

This repository contains the codes and data for the paper "Adaptive-Solver Framework for Dynamic Strategy Selection in Large Language Model Reasoning" which will soon be published in Information Processing & Management.

Notifications You must be signed in to change notification settings

john1226966735/Adaptive-Solver

Repository files navigation

Introduction

This repository is the official implementation of Adaptive-Solver (AS): Adaptive-Solver Framework for Dynamic Strategy Selection in Large Language Model Reasoning, which has been accepted by "Information Processing & Management".

Installation

pip install -r requirements.txt

Quick Start

We provide the solving records of each candidate solver on training dataset, as shown in "run_pipeline_configuration/solver2result/gsm8k/train". You need to fill in your own openai api-key in the beginning of "prediction_runner.py". Then, we execute the following two steps:

1. optimize phase: search for the optimal pipeline on the validation set

python run_pipeline_configuration/main_adapt_solvers_MSPD.py

2. evaluation phase: run our method AS_MSPD using the optimal pipeline on the test set

python run_evaluation/main_AUTO_gsm8k.py

About

This repository contains the codes and data for the paper "Adaptive-Solver Framework for Dynamic Strategy Selection in Large Language Model Reasoning" which will soon be published in Information Processing & Management.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages