Last 90 days · 2733 total
An arXiv paper examines the critical issue of web agent safety, specifically benchmarking their behavior when interacting with deceptive e-commerce i…
OriginalAn arXiv paper explores efficient on-device inference for Diffusion LLMs (dLLMs) using mobile NPUs, capitalizing on their parallel token denoising fo…
OriginalSalesforce has officially announced its definitive agreement to acquire Fin, indicating a strategic move in the AI industry.
OriginalAn article discussing Apple Foundation Models sparked comments and discussion on Hacker News, indicating community interest.
OriginalThe LocalLLaMA community expresses surprise and satisfaction with the significant improvements in the quality and effectiveness of KV quantization te…
OriginalAn arXiv paper investigates the run-to-run reliability and potential biases of "LLM-as-a-Judge" evaluation methods, despite their widespread use in r…
OriginalNovu Connect has launched on Product Hunt, offering a platform to "ship agents where your users already work," facilitating direct integration into e…
OriginalSlashy 是一款 AI 助手，可自动处理邮件工作。已在 Product Hunt 发布。
Originalllama.cpp 新版本 b9630 主要增加了对 Cohere2MoE 模型词汇表的支持（针对 TINY_AYA 模型）。这是 llama.cpp 的一次重要更新。
OriginalOllama v0.30.8 修复了启动时提供商选择问题，改进了提示缓存机制以更好地复用 KV 缓存，并提高了 MLX 推理的稳定性。
Original