The data used in the model includes prompts and intent categories derived from the GAIR-NLP/Auto-J scenario classification dataset. The responses from the model were evaluated pairwise using `openai/gpt-4-turbo-2024-04-09`. The embedding model was generated by first fine-tuning BAAI/bge-base-en-v1.5 with the intent categories from the dataset above, using contrastive learning with cosine similarity loss, and subsequently merging the resultant model with the base model at a 3:2 ratio.
Key takeaways:
- The pulze-intent-v0.1 is an Intent-tuned LLM router that selects the best LLM for a user query.
- The model uses a variety of models including claude-3-haiku-20240307, claude-3-opus-20240229, and gpt-4-turbo-2024-04-09 among others.
- Prompt and intent categories are derived from the GAIR-NLP/Auto-J scenario classification dataset.
- The embedding model was generated by first fine-tuning BAAI/bge-base-en-v1.5 with the intent categories from the dataset above, using contrastive learning with cosine similarity loss.