… View arXiv page View PDF GitHub 4 Add to collection Community Multi-agent systems MAS built on large language models are typically organized around roles, pipelines, and turn schedules, while the content that agents pass to one another is often left as unconstrained natural language. …
… OmniOpt thus supplies the research community with an operational coordinate system for selecting optimizers under explicit mechanism and objective assumptions, and charts a direction for the future development of the optimizer community. …
… Project Page: https://zhangrongxiang.github.io/leaptalk-page/ View arXiv page View PDF Project page GitHub 177 Add to collection Community Paper: https://arxiv.org/abs/2608.00079 Project Page: https://zhangrongxiang.github.io/leaptalk-page/ Code: https://github.com/zhangrongxiang/LeapTalk This is a… …
… With the Open ASR Leaderboard adding multilingual and long-audio tracks, it’s becoming one of the most reliable ways to gauge practical ASR performance today. why there is no models from China? · not intentional on our part and most models are contributed from the community! if you think an importa… …
… View arXiv page View PDF GitHub 2 Add to collection Community Pipeline parallelism is essential for training large neural networks, but existing schedules trade off throughput, memory, and optimization consistency. …
… View arXiv page View PDF GitHub 28 Add to collection Community We compare several recent recurrent linear-attention architectures —DeltaNet, Gated DeltaNet, Kimi Delta Attention, and Gated DeltaNet-2 — in a shared recurrent-memory notation and empirical setup. …
… View arXiv page View PDF Add to collection Community Adapting a foundation vision-language encoder to a specialized retrieval task creates a fundamental tradeoff: gains on the target distribution come at the cost of the foundation model's broad generalization, and fashion retrieval is a stringent i… …
… View arXiv page View PDF Project page Add to collection Community What happens when LLM agents are allowed to manage real capital, not just solve benchmark tasks? We study a 21-day deployment of 3,505 user-funded agents trading real ETH onchain. …
… View arXiv page View PDF GitHub 3 Add to collection Community 🤔 When does training a multi-agent LLM workflow end-to-end with RL actually help, and when does it quietly break? …
… View arXiv page View PDF GitHub 2 Add to collection Community PsychoSafe is a psychologically informed framework for making LLM refusals more supportive in high-risk situations, such as crises, coercion, or escalating harmful intent. …