131 个
akaOSS studio——五个面向人在回路 AI 测量与开发者工具的开源项目，Assist-Not-Complete 论文，可复现研究信息流，以及 HITL Kit 的 shadcn registry。站点 akaoss.dev。The akaOSS studio — five open-source projects for human-in-the-loop AI measurement and developer tooling, the Assist-Not-Complete paper, a reproducible research feed, and the HITL Kit shadcn registry. Live at akaoss.dev.
围绕 LLM 人类价值观与多元对齐的精选论文、基准、数据集与工具合集。A curated collection of papers, benchmarks, datasets, and tools on human values in LLMs and pluralistic alignment.
AI 安全测试——发现 LLM 应用、聊天机器人、AI Agent、MCP 服务器与 RAG 系统中的安全漏洞。对齐 OWASP（LLM、Agentic、MCP）与 NIST AI RMFAI security testing — find security vulnerabilities in LLM apps, chatbots, AI agents, MCP servers, and RAG systems. Mapped to OWASP (LLM, Agentic, MCP) and NIST AI RMF.
证据约束的代码诊断工作台：确定性静态分析拥有源码事实，LLM 只提假设，resolver/validator 决定能否绑定回真实操作——证据不足保持 unknown/abstained，从架构抑制幻觉。已在真实开源项目发现 bug（stunner #89 已被上游确认修复）。Python。
基于 Microsoft Agent Framework (.NET) 的 LLM Agent 三层评估：运行时护栏、PR 闸口、校准过的评审者。该框架可自我评估。Three-tier evaluation for LLM agents on Microsoft Agent Framework (.NET): runtime guardrails, a PR gate, and calibrated judges. The framework evaluates itself.
面向 AI agent 与 agentic workflow 的开源 MLflow 插件：涵盖 prompt、tool、skill、MCP server、RAG 知识库、评估、部署、可观测性以及 Aria copilot。Open-source MLflow plugin for AI agents and agentic workflows: prompts, tools, skills, MCP servers, RAG knowledge bases, evaluation, deployment, observability, and Aria copilot.
RAG 真的物有所值吗？ragornot 在真实 AWS Lambda + Bedrock 后端上，通过四种检索模式（Flat/BM25、Hierarchical、LLM-only、RAG）运行相同查询，并测量延迟、质量、成本和碳排放——用数据帮你决定是否使用 RAG。静态 Next.js 部署于 GitHub Pages。Does RAG actually earn its cost? ragornot runs the same query through four retrieval modes (Flat/BM25, Hierarchical, LLM-only, RAG) against a live AWS Lambda + Bedrock backend and measures latency, quality, cost, and carbon — so you can decide RAG-or-not with data. Static Next.js on GitHub Pages.
在同一 chunk 集合上对比 Lexical / Vector / Graph RAG，提供确定性评估、自我修正的 LangGraph agent 循环、PII 治理与 RAG 就绪度分析器。Lexical vs Vector vs Graph RAG over one identical chunk set, with deterministic evaluation, a self-correcting LangGraph agent loop, PII governance, and a RAG-readiness analyzer.
📚 构建并评估 RAG 流水线，实现数据导入、嵌入、检索与问答，并提供准确性与相关性指标。📚 Build and evaluate RAG pipelines to ingest, embed, retrieve, and answer questions with metrics for accuracy and relevance.
🔍 在 Weaviate 混合检索中评测 embedding 模型，基于自有数据或 MTEB 数据集评估 MRR@K、Hit@K、延迟与内存占用🔍 Benchmark embedding models in hybrid search with Weaviate. Evaluate MRR@K, Hit@K, latency, and memory using your data or MTEB datasets.
UMD 行星长期记忆算法、公式、基准、SDK 与可复现研究。UMD planetary long-term memory algorithms, formulas, benchmarks, SDKs, and reproducible research
大语言模型情境感知相关论文精选目录。Curated bibliography of papers on situational awareness in large language models
本人毕业学年所有天体物理课题汇总。毕业设计聚焦轨道转移优化，采用梯度下降法及 Markov Chain Monte Carlo 方法进行不确定性评估These are all the astrophysics projects in my final year. My final-year project focuses on orbital transfer optimisation using gradient descent and the Markov Chain Monte Carlo method for uncertainty evaluation.
🔒 模拟对抗行为以测试和强化 MCP 防御，无需真实利用或风险，确保稳健的安全评估🔒 Simulate adversarial behaviors to test and strengthen MCP defenses without real exploitation or risk, ensuring robust security evaluations.
精选的评分量表、检查清单、评分标准集、原则和评分指南汇总，用于对现代生成模型进行评分、排名、验证、过滤或训练。A curated list of rubrics, checklists, criteria sets, principles, and scoring guides used to score, rank, verify, filter, or train modern generative models.
确定性 SAT API、CNF 预处理工具与可复现研究标准。Deterministic SAT APIs, CNF preprocessing tools, and reproducible research standards.
硕士论文研究：使用 Terraform IaC、Atomic Red Team MITRE ATT&CK 模拟与 Python 统计分析，在 Microsoft Azure 中实证评估零信任架构（ZTA）与传统边界安全。Sheffield Hallam University 2026。MSc dissertation research: Empirical evaluation of Zero Trust Architecture (ZTA) vs conventional perimeter security in Microsoft Azure using Terraform IaC, Atomic Red Team MITRE ATT&CK simulations, and Python statistical analysis. Sheffield Hallam University 2026.
🔍 使用多种 prompt 技巧在多步数学问题上分析 Mistral-7B 模型的数学推理能力。🔍 Analyze the mathematical reasoning abilities of the Mistral-7B model using diverse prompting techniques on multi-step math problems.
一个 TypeScript LangGraph agent，通过搜索、摘要和 human approval 回答 research questions，使用 LangSmith 完整 trace 并可直接用于评估。A TypeScript LangGraph agent that answers research questions with search, summarization, and human approval – fully traced with LangSmith and ready for evaluation
通过实战教程构建有效的 AI 评测，使用真实 agent 提升生产系统技能🎯 Build effective AI evaluations through a hands-on tutorial, using a realistic agent to enhance your production system skills.
通过系统综述和患者病例场景分析识别潜在药物不良反应（ADR），包括对症状、用药史、剂量及时间线的评估，以确定可疑药物并判断所报告反应是否可能与药物相关。项目展示了基础的 Pharmacovigilance 能力identifying potential Adverse Drug Reactions through systematic review and analysis of patient case scenarios. It includes evaluation of symptoms, medication history, dosage, and timelines to determine the suspected drug and assess whether the reported reaction is potentially drug-related. The project demonstrates basic Pharmacovigilance
音乐研究语料库：面向音乐生成、MIR、感知与认知、AI 音乐系统、评估的 agentic 文献综述 — 覆盖符号、音频、多模态与 agentic 方向Music research corpus: agentic literature review of music generation, MIR, perception & cognition, AI music systems, evaluation — symbolic, audio, multimodal, agentic
围绕矩匹配情景生成与时序投资组合评估的可复现研究。Reproducible research on moment-matching scenario generation and temporal portfolio evaluation
论文《Physics-Informed Neural Networks and Neural Operators for PDEs: A Unified Taxonomy and Systematic Review》（Neural Networks）的补充材料：PRISMA 数据集、PDE 复杂度评分标准以及概念验证的统一基准。Supplementary materials for 'Physics-Informed Neural Networks and Neural Operators for PDEs: A Unified Taxonomy and Systematic Review' (Neural Networks): PRISMA datasets, PDE complexity rubric, and proof-of-concept unified benchmark
面向空间 LLM 与地理空间 AI Agent 的生产工程模式——验证优先架构、地理空间 RAG、prompt-to-spatial-SQL、tool routing 与评估。Production engineering patterns for spatial LLMs and geospatial AI agents — validation-first architecture, geospatial RAG, prompt-to-spatial-SQL, tool routing, and evaluation.
可复现研究基准实验：二维度量空间下的贪心 t-Spanner 构造算法A reproducible research benchmarking lab for Greedy t-Spanner Construction Algorithms in 2D Metric Spaces.
在 DABench 数据分析任务上对 DSPy RLM（Recursive Language Models）进行基准测试，使用自动评分实现基于代码的迭代评估Benchmark DSPy Recursive Language Models on DABench data analysis tasks with automated scoring for iterative code-based evaluation
论文一：NL2SQL 歧义检测——M1.5 分类法、确定性基线、人工金标评估与可复现研究制品Paper 1: NL2SQL Ambiguity Detection — M1.5 taxonomy, deterministic baseline, human-gold evaluation, reproducible research artifact
HateMirage——可解释的伪仇恨检测与多维推理，ICON 2026 共享任务。在 HateMirage 语料库（4,530 条带标注的伪仇恨评论）上进行目标识别 + 意图与隐含意义生成。HateMirage - Explainable Faux Hate Detection and Multi-Dimensional Reasoning - Shared Task @ ICON 2026. Target identification + Intent and Implication generation over the HateMirage corpus (4,530 annotated Faux Hate comments).
端到端 Python RAG 框架，包含文档摄取、语义搜索、对话式 AI、多用户检索、可观测性、结构化输出与评估。An end-to-end Python RAG framework featuring document ingestion, semantic search, conversational AI, multi-user retrieval, observability, structured outputs, and evaluation.
SURF2026 计算金融项目，研究感知自由边界的美式期权风险曲面，使用 CN/PSOR 基准、文献综述、实验报告以及 Codex 辅助的分步研究规划SURF2026 computational finance project on free-boundary-aware American option risk surfaces, using CN/PSOR benchmarks, literature review, experiment reports, and step-by-step Codex-assisted research planning.
Rust Graph Tracker：为 LLM 编程 Agent 提供数值与日期来源追踪，支持表达式求值与推导验证Rust Graph Tracker: numeric and date provenance tracking with expression evaluation and derivation verification for LLM coding agents
可复现研究：基于基准、统计模型与走步前向验证，检验可解释的市场信号能否预测 SPY 的五日方向。Reproducible research testing whether interpretable market signals can forecast SPY’s five-day direction using benchmarks, statistical models, and walk-forward validation.
受 MemGPT 启发、支持 OpenAI API 的长期记忆 Agent，具备记忆生命周期控制与评估能力。MemGPT-inspired long-term memory agent with OpenAI API support, memory lifecycle controls, and evaluation.
面向人脸域去风格化、结构条件化、评估与质量过滤的紧凑型可复现研究流水线。Compact reproducible research pipeline for face-domain destylization, structural conditioning, evaluation, and quality filtering.
面向 Agentic Dynamics 的实验工具：衡量 AI Agent 如何行为、恢复并产出已验证的结果Experimental instrument for Agentic Dynamics: measuring how AI agents behave, recover, and produce verified outcomes