Adaptive testing framework based on Item Response Theory for efficient LLM evaluation.

A robot vision system using gaze-integrated active vision for foveated processing and robot learning.

Shadcn components that look like Claude Code, Codex and Grok

Expert-annotated ground truth and code for an AI pentesting agent evaluation protocol.

A neuro-symbolic benchmark for constrained route planning in remote sensing.

An expert-grounded dynamic evaluation framework for agentic AI on open-ended professional tasks.

An agentic RL framework for autonomous travel planning using a hierarchical reward modeling system.

Google's lightweight open-weights AI model series.

A design-rules engine that prevents AI agents from building generic UI.

Multiscale Neural Operator leveraging LOD method for solving elliptic PDEs.

Tool-based visual evidence reasoning framework for Image Quality Assessment.

Compressed Activation REconstruction for Memory-Efficient LoRA fine-tuning.

A discrete-diffusion based speech recognition system using LoRA adapters.

Pipeline for constructing chronologically filtered datasets and training temporally valid LMs.

Code for evaluating automatic harness evolution vs test-time scaling for LLM agents.

Visual Interface Agent for Robot Control using frontier foundation models

Synonym-Coherent Language-Image Pretraining for Robust Open-Vocabulary Dense Perception.

Kinematics-agnostic human motion prediction via equivariant latent diffusion.

Lightweight hybrid model for automated ECG classification on resource-constrained devices.

Code, data, and model weights for Abstention-Aware Reinforcement Learning to mitigate LLM hallucinations.