Skip to main content
AI Engineer World's Fair (2024)

Deep Dive: Building SOTA Open Weights Tool Use: The Command R Family

PromptPanel

Head back to all of our AI Engineer World's Fair recaps

frame_0018.jpg

Sandra Kublik @itsSandraKublik / Cohere
Watch it on YouTube | AI.Engineer Talk Details

Sandra from Cohere gave a talk about their recent work on language models, focusing on their Command R family of models. She highlighted their progress in developing models optimized for retrieval-augmented generation (RAG) and tool use.

Key Points

frame_0191.jpg

  • Cohere released Command R and Command R Plus models in March 2024
  • These models excel at structured reasoning and are competitive with larger models like GPT-4 Turbo
  • Command R Plus quickly gained popularity in the open-source community
  • Cohere focused on optimizing for citations and reducing hallucinations
  • They open-sourced their UI toolkit for RAG applications in April 2024
  • Recent work has centered on improving tool use capabilities, especially for enterprise contexts

frame_0233.jpg

Technical Details

Model Design Decisions

frame_0424.jpg

  • Optimized for retrieval-augmented generation (RAG) and tool use
  • Addressed challenges like prompt sensitivity and overcoming model bias towards focusing on the beginning of documents
  • Improved ability to balance pre-trained knowledge with new information from prompts
  • Enhanced citation capabilities for better transparency and reduced hallucinations

frame_0466.jpg

Tool Use Capabilities

frame_0689.jpg

  • Developed both single-step and multi-step tool use functionality
  • Multi-step capability allows for sequential reasoning and error correction
  • Released a multi-step API that allows users to describe available tools and parameters
  • Model creates plans and adapts them based on tool outputs

frame_0824.jpg

Performance and Efficiency

frame_0529.jpg

frame_0899.jpg

  • Command R Plus is competitive with GPT-4 Turbo on complex reasoning benchmarks
    3-5 times cheaper to run than comparable models, improving scalability for production use

Demos

Sandra demonstrated two applications built with Cohere's models:

frame_0998.jpg

Complexity AI (cplxai): A generative search engine that provides answers grounded in multiple sources with clickable citations.

frame_1073.jpg

Internal RAG demo: Showed multi-step reasoning capabilities by asking the model to find the three largest companies by market cap, get employee counts, create a graph, and draft a tweet. The demo displayed each step of the model's thought process.

frame_1182.jpg

As well as usage with their Discord bot.

Resources