Tostino/Inkbot-13b-4k

TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Sep 21, 2023Architecture:Transformer0.0K Cold

Inkbot-13b-4k by Tostino is a 13 billion parameter conversational AI model designed to interpret and respond to structured prompts with or without contextual information. It excels in RAG-type queries, answering from context, and overriding memory, while aiming to be functional and token-efficient. The model features a unique structured prompt system, allowing dynamic dialogues based on metadata and user input, and supports a 4096 token context length. Its primary strength lies in its ability to handle diverse tasks like knowledge graph extraction, question answering, reasoning, translation, summarization, and creative writing through specific task options.

Loading preview...

Inkbot-13b-4k: Structured Conversational AI

Inkbot-13b-4k is a 13 billion parameter conversational AI model developed by Tostino, built for interpreting and responding to structured prompts. It is designed to provide accurate and meaningful interactions by leveraging a unique structured prompt system that allows for dynamic dialogues based on context, metadata, and user input. The model supports a 4096 token context length.

Key Capabilities & Features

  • Structured Prompt System: Utilizes specific <#tag#> formats for meta-data, system instructions, chat history, and user context, enabling precise control over interactions.
  • RAG-Type Query Excellence: Excels at answering queries from provided context and can temporarily override its memory with user_context.
  • Token Efficiency: Designed to be functional and less verbose, avoiding superfluous language to conserve tokens.
  • Diverse Task Options: Supports specialized tasks including general queries, knowledge_graph extraction, question_answer, reasoning, translation, summarization, and creative_writing.

Good For

  • Applications requiring precise control over AI responses through structured inputs.
  • Retrieval-Augmented Generation (RAG) scenarios where contextual information is paramount.
  • Developers needing a model that can perform specific functions like data extraction, summarization, or translation with explicit guidance.
  • Use cases where token efficiency and direct, less chatty responses are preferred.

Finetunes
1 models