Model Overview

Llama-3-8B-Stroganoff is an 8 billion parameter language model developed by HiroseKoichi, built upon the Llama-3 architecture. This model is a merge of several Llama-3-8B roleplay-focused models, including L3-8B-Stheno-v3.2, Hathor_Tahsin-L3-8B-v0.85, L3-8B-Lunar-Stheno, and badger-writer-llama-3-8b, with Meta-Llama-3-8B-Instruct as its base.

Key Differentiators & Capabilities

• Coherency-First Roleplay: Prioritizes logical and consistent responses, aiming to prevent illogical statements or contradictions that disrupt the roleplay flow.
• Optimized Response Lengths: Designed to provide concise responses (typically 10 sentences or fewer) while still offering detailed replies when contextually appropriate, avoiding excessive verbosity.
• Active Character Initiative: Characters generated by the model are noted for taking initiative and suggesting actions, enhancing dynamic interactions.
• Repetition Avoidance: Successfully tested to avoid common repetition traps in conversational AI, contributing to a more natural dialogue experience.
• Plaintext Formatting Preference: Optimized for plaintext + "quote" formatting, moving away from the often inconsistent action + "quote" style.
• 8K Context Window: Supports a context length of 8192 tokens, allowing for more extensive and detailed roleplay scenarios.

Ideal Use Cases

• Interactive Roleplaying: Excellent for users seeking a highly coherent and engaging roleplay partner.
• Narrative Generation: Suitable for generating dynamic and initiative-driven story elements.
• Conversational AI: Can be used in applications where consistent, non-repetitive, and appropriately detailed dialogue is crucial.