Ja-ck/Mistral-instruct-Y24-DPO

TEXT GENERATIONConcurrency Cost:1Model Size:7BQuant:FP8Ctx Length:8kLicense:apache-2.0Architecture:Transformer Open Weights Cold

Ja-ck/Mistral-instruct-Y24-DPO is a 7 billion parameter instruction-tuned causal language model based on the Mistral architecture. This model is specifically fine-tuned using Direct Preference Optimization (DPO) for enhanced instruction following. It features an 8192-token context length and is optimized for generating responses based on a clear question-answer prompt template.

Loading preview...

Ja-ck/Mistral-instruct-Y24-DPO Overview

This model, developed by Ja-ck, is a 7 billion parameter instruction-tuned language model built upon the Mistral architecture. It leverages Direct Preference Optimization (DPO) to refine its ability to follow instructions effectively, making it particularly adept at generating structured and relevant responses.

Key Capabilities

  • Instruction Following: Optimized through DPO to accurately interpret and respond to user instructions.
  • Context Handling: Supports a substantial context window of 8192 tokens, allowing for processing longer prompts and maintaining conversational coherence.
  • Structured Output: Designed to work efficiently with a specific ### 질문: {instruction}\n\n### 답변: {output} prompt template, facilitating clear question-answer interactions.

Good For

  • Applications requiring precise instruction adherence.
  • Question-answering systems where input and output follow a defined format.
  • Tasks benefiting from a 7B parameter model with an extended context length for detailed interactions.

Popular Sampler Settings

Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.

temperature
top_p
top_k
frequency_penalty
presence_penalty
repetition_penalty
min_p