UW-Madison-Lee-Lab/Qwen-PRM800K

TEXT GENERATIONPricing:Input $0.4 / Cached $0.08 / Output $0.8Concurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Feb 8, 2025License:apache-2.0Architecture:Transformer Open Weights Featherless Exclusive Cold

UW-Madison-Lee-Lab/Qwen-PRM800K is a 7.6 billion parameter causal language model, fine-tuned from Qwen/Qwen2.5-Math-7B-Instruct. This model specializes in mathematical reasoning and problem-solving, leveraging the PRM800K dataset for enhanced performance in generating and evaluating step-by-step solutions. With a context length of 32768 tokens, it is designed for tasks requiring detailed logical progression and reward signal generation for mathematical solutions.

Loading preview...

Model Overview

UW-Madison-Lee-Lab/Qwen-PRM800K is a 7.6 billion parameter language model developed by UW-Madison-Lee-Lab. It is a fine-tuned variant of the Qwen2.5-Math-7B-Instruct base model, specifically optimized for mathematical reasoning tasks.

Key Capabilities

  • Mathematical Problem Solving: The model is fine-tuned on the PRM800K dataset, which focuses on generating and evaluating detailed, step-by-step solutions to mathematical problems.
  • Solution Evaluation: It includes functionality to compute reward signals for candidate solution steps, indicating the correctness or quality of each step in a multi-step mathematical solution.
  • Extended Context: Supports a context length of 32768 tokens, allowing for processing and generating longer, more complex mathematical explanations and problem descriptions.

Use Cases

This model is particularly well-suited for applications requiring:

  • Automated grading or evaluation of mathematical solutions.
  • Generating detailed, logical explanations for mathematical problems.
  • Research into reinforcement learning from human feedback (RLHF) for mathematical reasoning, utilizing its reward signal generation capabilities.