Weyaxi/HelpSteer-filtered-Solar-Instruct

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:10.7BQuant:FP8Ctx Length:4kPublished:Jan 14, 2024License:apache-2.0Architecture:Transformer Open Weights Warm

Weyaxi/HelpSteer-filtered-Solar-Instruct is a 10.7 billion parameter instruction-tuned causal language model developed by Weyaxi. It is fine-tuned from Upstage's SOLAR-10.7B-Instruct-v1.0 using a filtered version of Nvidia's HelpSteer dataset. This model is optimized for following instructions and generating helpful assistant-style responses, leveraging its 4096-token context length for conversational tasks.

Loading preview...

Overview

Weyaxi/HelpSteer-filtered-Solar-Instruct is a 10.7 billion parameter instruction-tuned language model. It is built upon the robust foundation of upstage/SOLAR-10.7B-Instruct-v1.0 and further refined through fine-tuning.

Key Capabilities

  • Instruction Following: The model is specifically trained to understand and execute user instructions effectively.
  • Assistant-style Responses: It excels at generating helpful and coherent responses, making it suitable for conversational AI applications.
  • Dataset Utilization: Fine-tuned with a filtered version of Nvidia's HelpSteer dataset, enhancing its ability to provide guided and relevant outputs.

Good For

  • Chatbots and Virtual Assistants: Its instruction-following and assistant-style response generation make it well-suited for interactive applications.
  • General Instruction-Based Tasks: Can be applied to various tasks requiring the model to follow specific directives from users.

Prompt Template

The model utilizes a simple User-Assistant prompt template for interaction:

### User:
{user}

### Asistant:
{asistant}

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