Skip to content

THE AI WIREINTELLIGENCE THAT MATTERS

News Intelligence Research Models Markets Policy Briefings

ResearchPapers Technologies

The AI Wire — Intelligence for the global AI ecosystem·About·Privacy·Terms·Methodology·Contact·RSS

Last updated: —

THE AI WIREINTELLIGENCE THAT MATTERS

News Intelligence Research Models Markets Policy Briefings

ResearchPapers Technologies

Filtering by Model Efficiency✕

RESEARCH LEAD STORY

Function Lives Where Variance Doesn't: Task-Weighted Charts of a Language Model's Computation

How many dimensions does a language model's computation actually use? The question is ill-posed until one names a functional. Task-weighted charts make it well-posed: low-dimensional coordinate systems fit against a chosen functional of the representation, und

Read more Technologies

arXiv

|

17 Sept 2026

RESEARCH HIGHLIGHTS

Function Lives Where Variance Doesn't: Task-Weighted Charts of a Language Model's Computation

Who Judges Matters: Measuring Family-Conditioned Preference in LLM-as-Judge Panels

The Attention Within: Consensus Dynamics in Selective State Space Models

LATEST RESEARCH

Function Lives Where Variance Doesn't: Task-Weighted Charts of a Language Model's Computation

Who Judges Matters: Measuring Family-Conditioned Preference in LLM-as-Judge Panels

The Attention Within: Consensus Dynamics in Selective State Space Models

RESEARCH ATLAS

Explore all categories →

Agents122 papers

Multimodal Systems85 papers

Benchmarks68 papers

AI Safety37 papers

Retrieval20 papers

Model Efficiency7 papers

RESEARCH TIMELINE

SEPT 2026Function Lives Where Variance Doesn't: Task-Weighted Charts of a Language Model's Computation

SEPT 2026Who Judges Matters: Measuring Family-Conditioned Preference in LLM-as-Judge Panels

SEPT 2026The Attention Within: Consensus Dynamics in Selective State Space Models

SEPT 2026Can We Do Interpretable NLI with Graphs Based on Atomic Propositions?

SEPT 2026Cross-Block Conditioning in Deep Boltzmann Machines for Statistical Data Fusion

RECENT RESEARCH

Function Lives Where Variance Doesn't: Task-Weighted Charts of a Language Model's Computation

arXiv

Who Judges Matters: Measuring Family-Conditioned Preference in LLM-as-Judge Panels

arXiv

The Attention Within: Consensus Dynamics in Selective State Space Models

arXiv

Can We Do Interpretable NLI with Graphs Based on Atomic Propositions?

arXiv

Cross-Block Conditioning in Deep Boltzmann Machines for Statistical Data Fusion

arXiv

Browse every paper →

The AI Wire — Intelligence for the global AI ecosystem·About·Privacy·Terms·Methodology·Contact·RSS

Last updated: —