Fast all-in-one JavaScript runtime, package manager, and test runner written in Zig.

Framework for automating black-box, reconnaissance-driven pentesting of AI agents.

A framework for automatically formalizing geometry problems in native Mathlib using Lean.

A unified platform for comparing and debugging Graph RAG methods with visual retrieval tracking.

A Rotary Position Embedding variant with learnable head-specific rotation and scaling.

A multi-model LLM serving workload dataset and generator for benchmarking serving platforms.

An independent difficulty-stratified benchmark for open-vocabulary audio event grounding.

A curated list of resources and materials for AI in mathematical reasoning.

A formal framework for multi-level safety evaluation of foundation model-based embodied agents.

A framework enabling search agents to self-evolve their reasoning capabilities using external search engines.

A library for efficient, structured concurrency in OCaml.

Code and models for enhancing legal machine translation via SFT and RL.

Collaborative 3D object detection framework with 3D-aware fusion and refinement.

A controlled persona-matched benchmark for evaluating cross-agent campaign attribution in LLM agents.

A curated directory of legendary Hacker News links

A prototype system for agent provenance and sensemaking using structured provenance graphs.

A diagnostic framework for characterizing small LLMs for cybersecurity QA across vocabulary, knowledge, and contextualization.

Framework for learning representation programs to map raw documents into retrieval-optimized representations.

An empirical benchmark for evaluating the security of LLM agents in High-Performance Computing environments.

A spatio-temporal framework for enhanced traffic accident risk prediction using Mamba blocks and LSTM.