Jeva: a first-principles System One model architecture with null-aware, cacheable decisions
burny_tech · x · 2026-10-20
burnytech highlights the Jeva paper, a System One model architecture built from first principles for fast decision-making given a state and candidate options. Compared to conventional joint-encoding designs, Jeva adds four structural properties: (a) null-awareness with an explicit 'none of the above' outcome, (b) permutation invariance so candidate order doesn't change the choice, (c) natural support for variable-size candidate sets, and (d) cacheable option representations reusable across requests via an Option Memory block — promising more robust and efficient inference for decision tasks.
Related event: Jeva Paper Builds System One Decision Models from First Principles(2 posts)→
More from Research
- Foundation model trained on 40M face photos predicts cancer outcomes and biological age — EricTopol · 2026-10-20
- Cloudflare introduces clairic: a category-free model paradigm for extracting values from raw inputs — ritakozlov · 2026-10-20
- Why GRPO collapses on multi-turn agent training, and how to stabilize advantage estimation — burkov · 2026-10-20
- SJTU's MSAR Residual Architecture Delivers 1.58x Iso-Loss Training Speedup Over Standard Residuals — burny_tech · 2026-10-20
- From memory capacity to transformers: outer-product stores cap out near 2×key dimension — burny_tech · 2026-10-20
- UIUC paper: Fisher-selected 0.1% of parameters recovers 85% of full RL training gains — burny_tech · 2026-10-20