Davis Blalock on Muck Rack

Davis Blalock

As seen in: AI Supremacy, SemiAnalysis (Newsletter), Substack, Databricks

Get in touch with Davis

Contact Davis, search articles and posts on X, monitor coverage, and track replies from one place.

Learn more about Muck Rack

Actions

Share this page

Davis Blalock’s Biography

Read Full Bio →

Articles

$μ$nit Scaling: Simple and Scalable FP8 LLM Training

|
By Abhay Gupta, Mansheej Paul, Davis Blalock
| arxiv.org
arXiv:2502.05967 (cs) [Submitted on 9 Feb 2025 (v1), last revised 5 Jun 2025 (this version, v3)] View PDF HTML (experimental) Comments: Subjects: Machine Learning (cs.LG) Cite as: arXiv:2502.05967 [cs.LG] (or arXiv:2502.05967v3 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2502.05967 Submission history From: Abhay Gupta [ view email] [v1] Sun, 9 Feb 2025 17:31:09 UTC (6,226 KB) [v2] Sun, 2 Mar 2025 20:16:43 UTC (6,225 KB) [v3] Thu, 5 Jun 2025 08:10:47 UTC (6,487 KB) Bibliographic...
Open in Who Shared View Duplicates (225) Issue with article?

2024-8-25: Scaling curves for All of the Things

|
By Davis Blalock
| Substack
Good news: we got a bunch of important findings this week. Bad news: 40+ hours of work and 40+ hours of baby stuff doesn’t leave much time for papers, so we’ve got a shorter installment this week that’s more of a highlight reel. You know how there are a bunch of ways to get better results out of a fixed LLM by throwing more compute at the inference (e.g., majority voting across multiple generations)?
Open in Who Shared Issue with article?

Turbocharged Training: Optimizing the Databricks Mosaic AI stack with FP8

|
By Mihir Patel, Davis Blalock, Saaketh Narayan
| Databricks
At Databricks, we believe that the best companies in the world, in every sector, will have AI-powered systems that are trained and customized on their own proprietary data. Today's enterprises can maximize their competitive advantage by training their own AI models. We’re committed to providing the best platform for enterprises to train AI as quickly and cost-effectively as possible.
Open in Who Shared Issue with article?
See all 12 articles
 •  Privacy  •  Terms  •  Security and Legal  •   •   •   • 

© 2026 Muck Rack

SOC 2® Certified