featherless-ai/RWKV-mid-classifier
The featherless-ai/RWKV-mid-classifier is a 7.2 billion parameter RWKV model designed for classification tasks, leveraging the Simple Jev classifier API on Featherless. It specializes in converting text context into structured decisions, supporting choice, score, and noul question types. This model is optimized for applications like intent detection, content moderation, and routing, processing text and chat history inputs with a context length of 32768 tokens.
Loading preview...
Overview
The featherless-ai/RWKV-mid-classifier is a 7.2 billion parameter RWKV model integrated with the Simple Jev classifier API on Featherless. Its core function is to transform textual context, such as plain text or chat conversations, into structured, actionable decisions. This model is specifically designed for classification tasks, providing JSON outputs that can be directly consumed by applications.
Key Capabilities
- Structured Decision Making: Converts text input into structured JSON outputs for classification.
- Multiple Question Types: Supports
choice(selecting from named candidates),score(rating on an ordered rubric), andnoul(yes/no propositions with confidence scores). - Text and Chat History Support: Processes both raw text strings and conversational message histories as context.
- Independent Question Evaluation: Allows multiple questions to be evaluated against the same context, with each question assessed independently.
Good For
- Intent Detection: Identifying user intent in customer service or conversational AI.
- Content Moderation: Classifying content for appropriateness or policy adherence.
- Relevance Scoring: Determining the relevance of information to a query.
- Support Routing: Directing customer inquiries to the correct department or agent.
- Automating Decisions: Providing structured outputs for automated workflows based on textual input.