我的收藏提交工具
FOXAI SEARCH

全站搜索

新闻报道 462

用于LLM智能体一致性的递归语义分歧

arXiv:2610.33012v1 发布类型:新 摘要:具备工具调用能力的大型语言模型(LLM)智能体,在从相同初始状态出发的多次独立运行中,表现出行为不一致的现象,例如调用不同的工具、产生相互矛盾的输出,或采取截然不同的策略。这种行为上的波动会随时间累积,从而削弱部署的可靠性。现有的评估指标主要通过基于任务完成度的得分来衡量正确性,而未能捕捉智能体在多次执行中的行为一致性。此外,对智能体行为的分析要么停留在词元层面,借助概率分布进行评估;要么聚焦于单次响应层面,采用语义聚类方法,但两者均无法刻画不一致性如何在轨迹的后续步骤中进一步传播。我们提出了递归语义发散(RSD),一种无

arXiv cs.CL

b11927

metal : add GGML_OP_TOP_K_MASK (#31230) * metal : add GGML_OP_TOP_K_MASK Port the fused top-k mask from the CUDA backend: a grid-stride kernel fills the output with -inf, then one threadgroup per row copies the mask values of the kept cells over it. Function constants drop the score, blocks and extra paths that a given graph does not use. Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) * metal : clean up top-k mask kernel formatting Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11927/llama-b11927-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](

翻译服务暂未运行,当前显示原文。
llama.cpp 发布

b11907

models : replace ggml_tensor** out-params with output structs (#31142) The hyper-connection and DSA graph builders wrote their extra outputs (post / comb gates, carried top-k selections, inject gates) through ggml_tensor** out-params. Return them in a small struct next to the primary output instead. Assisted-by: dsh:llama.cpp/ggml-org/Qwen3.8-Flash-Next-GGUF (MacBook, M5 Max, 128GB) **Website:** - **Attestations:** - **macOS/iOS:** - [macOS Apple Silicon (arm64)](https://github.com/ggml-org/llama.cpp/releases/download/b11907/llama-b11907-bin-macos-arm64.tar.gz) - macOS Apple Silicon (arm64, KleidiAI enabled) [DISABLED](https://github.com/ggml-org/llama.cpp/pull/23780) - [macOS Intel (x64)](https://github.com/ggml-org/llama.cpp/releases/download/b11907/llama-b11907-bin-macos-x64.tar.gz) - [

翻译服务暂未运行,当前显示原文。
llama.cpp 发布

Release: v5.21.0

# Release v5.21.0 ## New Model additions ### MiniCPM-OCR MiniCPM-OCR is a document-understanding model from OpenBMB's MiniCPM-V family. It keeps the MiniCPM-V 4.6 architecture, which pairs a SigLIP vision encoder and a window-attention merger with a Qwen3.5 language model. It is specialized for reading text-dense images such as scanned pages, screenshots, receipts and tables. **Links:** [Documentation](https://huggingface.co/docs/transformers/main/en/model_doc/minicpm_ocr) | [Paper](https://huggingface.co/papers/2509.18154) * Feat test (#49920) by @guarin in [#49920](https://github.com/huggingface/transformers/pull/49920) ### Flint Flint is Recursive's mixture-of-experts language model. Compared to a standard SwiGLU MoE transformer such as Qwen3-MoE, it uses hybrid attention (sliding-windo

翻译服务暂未运行,当前显示原文。
变压器发布

面向大语言模型可靠性检测的表征签名自适应迁移

arXiv:2610.33442v1 发布类型:新 文摘:据报道,大型语言模型(LLMs)容易出现包括幻觉和推理错误在内的可靠性问题。现有的质量检测器能够基于内部表征以相当高的准确率预测不可靠的输出。然而,这些检测器的训练通常需要大量特定领域的标注数据,这限制了它们在新领域的适用性。我们提出通过在相关领域之间迁移表征签名来应对这一挑战。具体而言,我们构建了一个统计框架,首先为每个领域提取低维表征签名,然后利用交叉拟合学习各领域特定预测器的组合权重。这些权重会根据源领域各自的预测相关性进行自适应调整。

arXiv cs.CL