2 个仓库
Tools that validate the correctness and reliability of a model's function calling and API triggering capabilities.
Distinct from Function Calling Interfaces: Focuses on the evaluation/verification of the calls, whereas the parent focuses on the interface that enables the calls.
Explore 2 awesome GitHub repositories matching artificial intelligence & ml · Function Call Verifiers. Refine with filters or upvote what's useful.
Giskard 是一个用于大语言模型和 AI 智能体的评估框架、测试库及质量监控系统。它作为量化模型性能和可靠性的工具包,为验证检索增强生成(RAG)流水线提供了专门的功能。 该项目通过自动化的红队测试工具和安全扫描器脱颖而出,旨在识别漏洞、提示词注入和安全风险。它利用对抗性探测和合成边缘案例生成来量化模型的鲁棒性并检测信息泄露。 该平台涵盖了广泛的功能,包括事实准确性与幻觉检测、推理与逻辑基准测试以及偏见检测。它提供了回归测试、RAG 组件评估以及从知识库自动生成测试用例的工具。 该系统包含协作团队工作区管理、基于角色的访问控制以及用于监控性能漂移的定时评估流水线等功能。
Validates the ability to trigger correct functions and APIs across multiple languages, including parallel execution.
ACI is a tool-calling platform and centralized system for managing and executing external service operations and custom scripts for agentic workflows. It functions as a unified Model Context Protocol server that enables AI agents and IDEs to dynamically discover and execute diverse toolsets. The platform distinguishes itself through a natural language capability index and intent matching to search for available tools based on task requirements. It provides an external service authenticator and account linking via OAuth-based credential management to permit secure tool execution on behalf of u
Translates natural language intent and agent requests into specific executable function calls for external services.