Neelectric/Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.20
Neelectric/Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.20 is an 8 billion parameter instruction-tuned Llama-3.1 model developed by Neelectric. Fine-tuned on a specialized mathematical dataset, it is optimized for mathematical reasoning and problem-solving tasks. This model leverages a 32768 token context length, making it suitable for complex mathematical queries and detailed explanations.
Loading preview...
Overview
This model, Neelectric/Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.20, is an 8 billion parameter instruction-tuned variant of the meta-llama/Llama-3.1-8B-Instruct base model. Developed by Neelectric, it has been specifically fine-tuned using Supervised Fine-Tuning (SFT) on the Neelectric/OpenR1-Math-220k_all_Llama3_4096toks dataset, which focuses on mathematical content.
Key Capabilities
- Enhanced Mathematical Reasoning: Optimized for understanding and solving mathematical problems due to its specialized training data.
- Instruction Following: Benefits from the Llama-3.1-Instruct base, providing strong instruction-following capabilities.
- Extended Context: Supports a context length of 32768 tokens, allowing for processing and generating longer, more complex mathematical discussions or problem descriptions.
Training Details
The model was trained using the TRL (Transformers Reinforcement Learning) library. The training process involved SFT on a dataset comprising 220,000 mathematical entries, specifically tailored for Llama 3 with a 4096-token context. This targeted fine-tuning aims to improve its performance on math-related tasks compared to its base model.
Good For
- Applications requiring robust mathematical problem-solving.
- Generating detailed explanations for mathematical concepts.
- Educational tools focused on mathematics.
- Research in mathematical reasoning with large language models.