TroyDoesAI/BlackSheep-X-Dolphin

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:8kPublished:Jul 14, 2024License:cc-by-nc-2.0Architecture:Transformer0.0K Open Weights Warm

TroyDoesAI/BlackSheep-X-Dolphin is a merged language model combining TroyDoesAI/BlackSheep with cognitivecomputations/dolphin-2.9.3-llama-3-8b. This model is noted for its strong performance, having held a top position on UGI Leaderboards for 8B models. It is designed for general conversational AI and instruction-following tasks, leveraging the strengths of its constituent base models.

Loading preview...

TroyDoesAI/BlackSheep-X-Dolphin: A High-Performing Merged Model

TroyDoesAI/BlackSheep-X-Dolphin is a powerful merged language model, built upon the foundation of TroyDoesAI/BlackSheep and cognitivecomputations/dolphin-2.9.3-llama-3-8b. This strategic merge aims to combine the strengths of both base models, resulting in a highly capable instruction-following AI.

Key Capabilities

  • Strong Instruction Following: Inherits and enhances the instruction-following capabilities from its Dolphin component.
  • Optimized for Conversational AI: Designed to excel in interactive dialogue and general-purpose chat scenarios.
  • Competitive Performance: The model has demonstrated significant performance, achieving and maintaining a #1 spot on the UGI Leaderboards for 8B models, indicating its efficiency and effectiveness in various benchmarks.

Good For

  • General AI Applications: Suitable for a wide range of tasks requiring robust language understanding and generation.
  • Research and Development: Provides a strong base for further fine-tuning or experimentation in the 8B parameter class.
  • Interactive Agents: Its conversational prowess makes it ideal for chatbots, virtual assistants, and other interactive AI systems.

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