Neelectric/Llama-3.1-8B-Instruct_SFT_Math-220kfisher_v00.02

TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Jan 20, 2026Architecture:Transformer Cold

Neelectric/Llama-3.1-8B-Instruct_SFT_Math-220kfisher_v00.02 is an 8 billion parameter instruction-tuned causal language model developed by Neelectric. It is a fine-tuned version of Meta's Llama-3.1-8B-Instruct, specifically optimized for mathematical reasoning and problem-solving. This model leverages a specialized dataset for mathematical tasks, making it suitable for applications requiring robust numerical and logical capabilities.

Loading preview...

Neelectric/Llama-3.1-8B-Instruct_SFT_Math-220kfisher_v00.02 Overview

This model is an 8 billion parameter instruction-tuned language model developed by Neelectric. It is built upon the Meta Llama-3.1-8B-Instruct architecture, fine-tuned using Supervised Fine-Tuning (SFT) with the Neelectric/OpenR1-Math-220k_extended_Llama3_4096toks dataset. The training process utilized the TRL framework, indicating a focus on enhancing instruction-following capabilities, particularly in mathematical contexts.

Key Capabilities

  • Enhanced Mathematical Reasoning: Specialized fine-tuning on a large mathematical dataset aims to improve the model's ability to understand and solve mathematical problems.
  • Instruction Following: As an instruction-tuned model, it is designed to respond effectively to user prompts and instructions.
  • Llama 3.1 Base: Benefits from the strong foundational capabilities of the Llama 3.1 series.

Good For

  • Applications requiring mathematical problem-solving or numerical reasoning.
  • Tasks where a robust instruction-following model with a focus on quantitative understanding is beneficial.
  • Developers looking for a Llama 3.1-based model with specialized mathematical SFT.