Norquinal/PetrolLM-Claude-Chat Overview

Norquinal/PetrolLM-Claude-Chat is a 7 billion parameter model built upon the Mistral-7B-claude-chat base, further specialized by the application of the PetrolLoRA. This LoRA was trained on a unique dataset of approximately 2800 samples, meticulously curated from sources like AICG Logs, PygmalionAI/PIPPA, and Squish42/bluemoon-fandom-1-1-rp-cleaned, with additional content from OpenLeecher/Teatime. The training data was back-filled or converted using GPT-4/GPT-3.5-turbo-16k to fit a specific roleplay-oriented prompt format.

Key Capabilities

• Specialized Roleplay Generation: Optimized for creating detailed and immersive character interactions.
• Structured Prompting: Utilizes a defined prompt format including style, characters, summary, and chat_history to guide generation.
• Descriptive Output: Designed to produce rich, multi-paragraph responses with vivid details about character actions, emotions, and environments.

Good For

• Text Generation Web UI: Compatible with transformers (bleeding-edge version or model_type change to llama).
• SillyTavern UI: Enhanced performance with specific Last Output Sequence prompts to encourage detailed and engaging replies, with a recommended system note for optimal results.

This model is particularly well-suited for applications requiring high-quality, narrative-driven conversational AI, especially within roleplaying or creative writing contexts.