neuron-explorer
Steering-first neuron explorer for Qwen3.5-4B: push every MLP neuron, keep every rollout, browse the results
翻译准备中
AI systems engineer who designs and ships production AI end to end - RAG pipelines, agentic assistants, measured retrieval quality, and multi-provider LLM infrastructure, with hands-on QLoRA fine-tuni
翻译准备中
开发工具Python · TypeScriptGitHub @architecture-zero在线运营者未确认
试用过了吗? 点一下就好,无需登录。
个人 GitHub 账号下公开仓库的所有者可以登录并验证所有权,自行管理介绍,无需管理员批准。
登录并生成验证文件,在 30 分钟内将其提交到此仓库的默认分支。其他分支或分叉仓库中的文件无效。然后在下方检查文件,即可无需管理员批准直接管理项目。请保留文件以维持管理权限。
如需登录,将前往登录页面。登录后请回到此处,再次点击所选的验证按钮。
在项目文件夹中运行 /nomorevibe verify,即可验证所有权并自行管理介绍。
一分钟完成验证A self-hosted AI assistant platform where trust is measured, not claimed - hybrid retrieval behind a reranker seam, clearance enforced identically at four surfaces including both sides of the federation seam, an ingest-time injection gate, and judged evals published live on a trust panel.
翻译准备中
取自仓库 README
正在加载完整 README。 以下为已保存的摘录。
A self-hosted AI assistant platform where trust is measured, not claimed.
Most "AI systems" are a text box, an API key, and a prompt. This is the
rest: hybrid retrieval with a cross-encoder reranker behind a swappable
provider seam, tiered access control enforced at four surfaces, an
ingest-time injection gate that quarantines hostile content before it can
reach an answer, and a judged evaluation harness that scores the system's
correctness, groundedness, freshness, and honesty - then publishes those
numbers on a public trust panel derived live from stored runs.查看完整 README ↗目前无法加载完整 README。请通过原文链接阅读。 以下为已保存的摘录。
A self-hosted AI assistant platform where trust is measured, not claimed.
Most "AI systems" are a text box, an API key, and a prompt. This is the
rest: hybrid retrieval with a cross-encoder reranker behind a swappable
provider seam, tiered access control enforced at four surfaces, an
ingest-time injection gate that quarantines hostile content before it can
reach an answer, and a judged evaluation harness that scores the system's
correctness, groundedness, freshness, and honesty - then publishes those
numbers on a public trust panel derived live from stored runs.查看完整 README ↗