distil labs published a blog post titled the three ways a small model actually enters a production system, and how to tell which one you are looking at.
Public source
Publisher name
Public post
New on the blog: the three ways a small model actually enters a production system, and how to tell which one you are looking at. Choosing the task is the easy part. Plac…
Company
distil labs
Replace LLMs with custom SLMs. Today. Faster, cheaper, just as accurate.
- Industry
- Software Development
- Location
- Berlin, DE
- Company size
- 11–50 employees
About distil labs
Replace LLMs with custom SLMs. Today. Once you start scaling your AI product, efficiency begins to matter quicker than you expect. Higher margins mean faster growth and better valuation multiples, becoming as big of a lever as your next feature. distil labs replaces your costly LLM with task-specific SLMs, lowering costs by 50-90% and cutting latency. Build your custom models in hours, no data labelling required. Simply upload your model traces and get started right away.
See moreLatest activity
Latest activity from distil labs
4 signals
Presence & Recognition
distil labs is attending the HumanX conference in Amsterdam from 22 to 24 September.
Presence & Recognition
distil labs is attending the HumanX conference in Amsterdam from September 22 to 24.
Research & Knowledge
distil labs published a blog post on the three ways to fit a small, purpose-trained model into an existing system.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
Cohere
Cohere published a blog post making the case for model-vendor FDEs and how they can bring product and problem-solving together.
Research & Knowledge
Foundation AI
Foundation AI published a blog post detailing the real work involved in evaluating and calibrating enterprise AI models to ensure they deliver correct answers without error or flag raising.
Research & Knowledge
Together AI
Together AI published research on QLoRA to compress the base model to 4 bits and reduce memory usage to 1/4, allowing the base model plus 16-bit adapters to fit on a smaller GPU.
Research & Knowledge
Bespoke Labs
Bespoke Labs published a blog post detailing how they performed SFT and RL to improve the performance of an open model on specific code repositories.
Research & Knowledge
Liquid Technologies