KaliBench: A Fine-Grained Benchmark for Cybersecurity Tool Use on Kali Linux with Runtime-Free Verifiable Rewards· KaliBench:Kali Linux 网络安全工具使用微基准测试
LLMs are increasingly applied to cybersecurity workflows, where they are expected to trans…
LLMs are increasingly applied to cybersecurity workflows, where they are expected to trans…
What makes great scientists great? Even as AI systems start to make progress on open probl…
We show that multimodal models possess strong reasoning abilities and that an appropriate …
Discrete diffusion language models offer a compelling alternative to autoregressive genera…
Coding agents solve repository-level software engineering tasks through long trajectories …
Large language model (LLM) agents increasingly rely on persistent external sources to solv…
Keyword-matching benchmarks can credit small models for tool use they never perform. We do…
A comparative explainability framework is presented to audit DeBERTa-v3 under zero-shot cl…
Privacy-preserving machine learning presents significant deployment challenges on the clou…
The study of ceramic materials constitutes a cornerstone of archaeological research, yet t…
Memory-augmented large language models must decide which memories to retain, and recent sy…
As Large Language Models (LLMs) increasingly serve as foundational reasoning engines, thei…
When a SCADA alarm is raised in a water distribution network, operators must decide quickl…
Large language model (LLM) agents increasingly combine reasoning, tool use, and action, bu…
AI agents can each make locally valid decisions yet jointly produce an invalid result. We …
While Large Language Models (LLMs) advance 3D indoor scene synthesis, current pipelines fa…
The rapid maturation of artificial intelligence (AI) and machine learning (ML) has catalyz…
A multi-LLM \emph{council} lets several large language models (LLMs) deliberate on a quest…
Small open-weight models (2-9B) run on ordinary laptops, but under cloud-scale agent harne…
AI systems increasingly produce outputs from confidential data, such as a fitness-for-duty…
Emergency department (ED) triage is a high-stakes prioritization task in which demographic…
Large Language Models (LLMs) have demonstrated remarkable fluency across many tasks but re…
Natural language descriptions can provide rich semantic representations of audio-visual ur…
Our research investigates how two adaptive AI methods, evolutionary transfer learning and …
A time series world model (TSWM) predicts a controlled system's state from its observed hi…
A simple Windows AI agent that runs locally with a single .exe file. No command line or co…
Open-source AI workspace for humans and agents — SuperCompress as the core context layer.
Бот, который сам играет в блок в DDNet · A bot that plays block in DDNet by itself. Window…
Turn token activity into compute. A compute-access protocol designed for Robinhood Chain: …
An open source, self-hosted, safety-focused AI workbench
We created a list of the most notable AI agents that can live in your text messages, from …
岗位JD甩了篇技术报告
又咋啦。。。
Meta wants Muse in your TV and your toaster, so it's giving the code away for free.
Sean Parker, who once taught the music industry what asking for forgiveness looks like, is…
Apple says it will add new controls around macOS’s Full Disk Access permission, warning th…
This week, the White House got nearly every major tech CEO in one room — Zuckerberg, Bezos…
This week, the White House got nearly every major tech CEO in one room — Zuckerberg, Bezos…
With all the talk about how AI might one day ill us all, it's easy to forget that AI has a…
Learn how startups can choose GPT-6 models, tune reasoning effort, improve prompts and ski…
"There is an ontological difference, even before an aesthetic one, between art and what a …
Here are Google’s latest AI updates from September 2026
Chatham Financial uses Codex and GPT-5.6 to build technology and redesign workflows, cutti…

YouTube · NVIDIA