AI Wiki — Knowledge Base

Introducing 𝐒𝐩𝐞𝐜𝐮𝐥𝐚𝐭𝐢𝐯𝐞 𝐃𝐞𝐜𝐨𝐝𝐢𝐧𝐠: 𝐇𝐨𝐰 𝐈𝐭 𝐄𝐯𝐨𝐥𝐯𝐞𝐝, 𝐖𝐡𝐞𝐧 𝐈𝐭 𝐒𝐭𝐚𝐲𝐬 𝐋𝐨𝐬𝐬𝐥𝐞𝐬𝐬, 𝐚𝐧𝐝 𝐖𝐡𝐚𝐭'…

Introducing 𝐒𝐩𝐞𝐜𝐮𝐥𝐚𝐭𝐢𝐯𝐞 𝐃𝐞𝐜𝐨𝐝𝐢𝐧𝐠: 𝐇𝐨𝐰 𝐈𝐭 𝐄𝐯𝐨𝐥𝐯𝐞𝐝, 𝐖𝐡𝐞𝐧 𝐈𝐭 𝐒𝐭𝐚𝐲𝐬 𝐋𝐨𝐬𝐬𝐥𝐞𝐬𝐬, 𝐚𝐧𝐝 𝐖𝐡𝐚𝐭'𝐬 𝐍𝐞𝐱𝐭. An interactive tutorial @Madisonkanna and I built for the NeurIPS Education Track. Every LLM you use generate

tutorialsswyx--x

6 Sept 2026

Search: “inference”

CrossDepth: Geometry-Constrained Attention for Generalizable Multi-View Surround Depth Estimation

Distilled Continuous Diffusion Language Models Can Write Code in Few Steps---or One

Encore: Infinite Audio-Video Generation with Adaptive Signal Routing

ExLlamaV3 is underrated

From Deep to Shallow: Unconstrained and Efficient Layer Merging Strategy

From Interpretability Methods to Interpretable Models

GEPARD - Generative, Prosody-aware, Autoregressive text-to-speech model for Realtime Dialogue

HaptiNet: Networked Haptic Robots Enable Physical Co-presence in Geographically-Unconstrained Rehabilitation

Harnessing the Reasoning Economy: A Survey of Efficient Reasoning for Large Language Models

Human-Human & Human-Robot Interaction Transformer (H2INT) for Robot Navigation in Dense and Uncertain Crowds

IPGeoAI: Transformer-Based Geolocation with LLM Semantic Fusion

Iris: Climbing to the Search Frontier

KVMem: Virtualizing Million-Token Agent Workspaces on a Consumer GPU

LayerStoRm open-source expert streaming: 1M context GLM-5.3-Flash [UD-Q4_K_XL] at 24.5 tok/s @8k on just 2× RTX 5090 + 2× RTX 5080 (186 GiB MoE on 96 GB VRAM)

LookThere! Sparse Vision by Reinforced Selection

Low-Latency Spell Correction for Japanese Music Search Queries

MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought Compression

Measuring proximity to standard planes during fetal brain ultrasound scanning

MINT: A Unified Model for World-Space Camera and Hand Motion Estimation from Scalable Egocentric Pipeline Supervision

Mitra-v2 Technical Report

MomentQuant: an even more minimalist interval method with linear time complexity for time series classification

Motion-Omni: End-to-End Joint Speech and Full-Body Motion for Spoken Dialogue

One Diffusion Model, Two Roles: Guided Trajectory Planning and Safety-Critical Scenario Generation in Closed-Loop Simulation

Orientation-Robust Latent Motion Trajectory Learning for Annotation-free Cardiac Phase Detection in Fetal Echocardiography

ParaRNN: Parallel Training for Non-Linear RNNs (sLSTM, RWKV-7, CfC, Titans) via Triton Newton Scans [P]

PetQA: Benchmarking Veterinary Knowledge and Clinical Reasoning

Pitch-class Steering for Diffusion-based Music Generation via Latent-space Probes

Recurrence Is Not Enough: Causally Validating Multilingual SAE Translation Features in Gemma 2 and 3

RefDiT: Local Attribute Guidance in Reference-Based Image Generation

Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models

RL-VLA^3: A Flexible and Asynchronous Reinforcement Learning Framework for VLA Training

Scalable Detection of Fossil Palynomorphs in Multifocal Digital Microscopy Images

SimFuse3D: Source-Guided Target Simulation and Confidence-Guided Multi-Stage Localization Reweighting for Cross-Platform 3D Object Detection

Spectral-Target Physical Latent Structuring for JEPA-Style World Models

Step Back to Move Forward: Reflection-Aware Preference Optimization for Visual Generation

⚠️ The most important question hardly anyone is asking is whether Jensen meant 100K+ GPUs, or 100k+ NVLink72 server racks (which contain 72 …

Too Rare to Learn: Prescribed Cyclone Tracks Degrade a Bay of Bengal Ocean Emulator

Towards Efficient Parametric State Estimation in Circulating Fuel Reactors with Shallow Recurrent Decoder Networks

Towards Neuro-Symbolic Procedural Reasoning for Long-Horizon Vision-Language-Action Manipulation

Towards Understanding Pause Token Fine-Tuning Dynamics: A Mode Retention Perspective

Tracing Audio Grounding and Answer Selection in Audio LLMs

Training-Free Halving of Activated Experts in Fine-Grained Mixture-of-Experts Models

UniMate: One Unified Model to Animate Diverse Skeletons

We are excited to bring the first open benchmarking of Google's TPUs to the world Running every day, on many models + scenarios $/token is b…

When Does an Interpretation Count as Established? The Formation, Evaluation, and Responsibility of Interpretation in Generative AI

Where to Look Matters: Learning Influential Views for VLM-based 3D Visual Grounding

“Currently I believe that no lab has solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed…

Introducing 𝐒𝐩𝐞𝐜𝐮𝐥𝐚𝐭𝐢𝐯𝐞 𝐃𝐞𝐜𝐨𝐝𝐢𝐧𝐠: 𝐇𝐨𝐰 𝐈𝐭 𝐄𝐯𝐨𝐥𝐯𝐞𝐝, 𝐖𝐡𝐞𝐧 𝐈𝐭 𝐒𝐭𝐚𝐲𝐬 𝐋𝐨𝐬𝐬𝐥𝐞𝐬𝐬, 𝐚𝐧𝐝 𝐖𝐡𝐚𝐭'…