A method for conditioning LLMs with personas that share user concerns to enhance rapport.

Research identifying limits of persona-conditioned simulation in Large Language Models.

X+Slides

★ 65

A benchmark for audience-conditioned slide generation using source-grounded probes.

A method for evaluating agentic systems by comparing temporal preferences over trajectories.

Study and framework for rectifying LLM-generated code using execution feedback loops.

Novel approach for integrating diverse feature extractors and graph representations in GNNs.

Self-supervised GNN framework for network intrusion detection using timestamps.

A self-evolving framework for zero-shot object goal navigation using agentic rule memory.

DRFLOW

★ 65

Deep research benchmark for personalized workflow prediction.

An assessment of AI's ability to solve research-level mathematics problems.

A structured analysis and taxonomy of security challenges in long-horizon agentic AI systems.

An economic foundation for the valuation and pricing of AI agents in human-AI workflows.

A terminal-based user interface for managing Azure DevOps projects and pipelines.

MR-GVNO

★ 65

Geometry-Aware Variational Physics-Informed Neural Operator for irregular plate problems.

Nxui

★ 65

Copy-paste animated UI components for Vue

Homebox

★ 65

An open-source inventory and organization system built for home users.

A public benchmark focused on GPU kernel optimization and LLM coding capabilities.

Beagle

★ 65

A tool for managing Git references via URIs.

An open-source tool to detect songs generated using the Treblo AI music system.

Audio demos and project page for studio-quality generative speech enhancement.