Johns Hopkins analysis of 1,100+ models claims all deep networks converge to one 16-D subspace
evolvingstuff · x · 2026-09-16
A widely shared thread claims Johns Hopkins researchers analyzed 1,100+ trained models across major architectures and proposed the 'Universal Weight Subspace Hypothesis': regardless of initialization, data, or task, deep neural networks collapse into the same low-dimensional geometric footprint, with most of a model's 'intelligence' living in a tiny shared spectral subspace controlled by a few principal directions. Note: thread-style retelling; verify against the actual paper.
More from Research
- Therna Bio opens Chronos platform, releases gene expression datasets with millions of measurements — AllThingsApx · 2026-09-16
- GlossoGen: open-source platform shows AI agents spontaneously inventing new languages under pressure — EliasEskin · 2026-09-16
- Paper: upsampling alignment discourse in pretraining cuts misalignment from 45% to 9% — TuhinChakr · 2026-09-16
- One researcher taught world model Odyssey-3 to drive in India with just 20 hours of data — soleio · 2026-09-16
- Is there a real hardware lottery in AI? Transformers may have won by fitting GPUs — prateekj · 2026-09-16
- Every tests a probability-output foundation model: 25x faster, ~600x cheaper than Fable — danshipper · 2026-09-16