A theoretical framework for analyzing and extending LLM policy optimization algorithms.

Paradigm for executable memory for personalized AI agents using typed Python objects.

An Adversarial Robustness Benchmark for World Models in Continuous Control.

A loss-aware many-shot merging method for mitigating task interference in LLMs.

Trust-native agentic routing infrastructure for secure AI agent communication and billing.

Strategic Trajectory Reasoning via Discriminative Estimation for Verifiable RL.

A benchmark of 1,800 task instances for process-level analysis and evaluation of web agents.

A benchmark of recurrent logic puzzles with BFS-optimal trajectories for evaluating LLM reasoning.

A Python package for implementing uncertainty-aware governance and oversight in delegated AI systems.

Benchmark for solving irregular Time Series Question Answering via tool-grounded reasoning.

Framework for optimizing agentic reasoning via semantic information gain rewards.

A reasoning-oriented post-training dataset for Large Audio-Language Models containing ~191k samples.

A file-based collaboration protocol for heterogeneous LLM agent collaboration.

Code for learning continuous-time physical prototypes for link failure detection.

A distributed content-addressable network for decentralized AI and data sharing.

A frequency-aware CNN-Transformer framework for Auditory Attention Decoding.

A conditioning-aware framework for stabilizing inversion and attention in diffusion image editing.

Multi-Tier LLM Inference Middleware with Dual-Channel HPC Token Streaming

Benchmark for measuring social engineering vulnerabilities in LLM-based code review agents.

A dataset of 342 modules with hierarchy metadata for hardware verification.