Find a story
Search Spins
Search titles, summaries, and missing voices across published articles — press releases, announcements, and media coverage.
0 results for “model capacity”
Scaling Laws, Tabular Data and Actuarial Ratemaking Models
A research paper investigates whether deep learning scaling laws apply to actuarial ratemaking using real motor insurance data, finding that data scaling behavior varies significantly across model families — with TabM outperforming Transformers and MLPs — and that architectural design and loss objectives matter more than raw parameter count.
Sep 4, 2026
EntropyMoE: Entropy-Aware Sparse Expert Routing for Tokenizer-Free LLMs
EntropyMoE is a new Mixture-of-Experts architecture for byte-level, tokenizer-free LLMs that routes computation per dynamic byte patch using entropy as a routing signal, improving compression efficiency without sacrificing downstream task accuracy.
Aug 10, 2026