ai-hallucinations

A lightweight framework for fact-checking AI-generated content. 解决由模型幻觉产生重大问题的轻量级框架。

Noetic Geodesic Framework for AI Reasoning

Prompts that Can Suppress AI Hallucinations 可以抑制AI幻觉的提示词

Deliberately broken, insecure, non‑compliant code and GDPR/NIS2/CRA violations examples for training AI.

Framework structures causes for AI hallucinations and provides countermeasures

Interactive daily quiz to detect and fix AI hallucinations with shareable mission reports

The Open Hallucination Index is an open-source initiative dedicated to enhancing AI safety by providing a robust toolkit for measuring factual consistency and mitigating generation errors in modern Generative AI architectures.

An open-source, community-driven repository documenting real-world AI failures, including hallucinations, logic errors, mathematical inaccuracies, legal misinterpretations, and security vulnerabilities.

Trust layer for domain-specific AI

Verify AI-generated code before you commit — 3 seconds, zero config. ⭐ Catch hallucinations in Claude, Copilot, Cursor, ChatGPT code. Open source. Sponsor ❤️

Stop AI Hallucinations & Flag Disinformation. 🛡️ Official Veritier API samples for real-time fact-checking, automated claim extraction, and Truth Firewall integration. Build reliable AI applications and secure your content entry-points with Python, JavaScript, and MCP examples. 🚀

Rispondere senza sapere: Discursive Strategies of Large Language Models Under Information Scarcity. The Case of Fanciulla di Vagli - Research paper on Large Language Models, information scarcity, hallucinations, source attribution, and discursive strategies. Case study: Fanciulla di Vagli. (Zenodo DOI: 10.5281/zenodo.19812343)

Low-overhead execution-finality reference implementation for AI/GPU environments: bounded non-bearer authority, atomic extraction-state control, Finality Sink enforcement, RATS attestation mapping, and performance-aware paths for multi-GPU, streaming, batching, confidential computing, DPU/SmartNIC, firmware, and silicon.

Logic-first diagnostic engines for emotion, clarity, and control.

Truth-Grounded ATS Resume Tailoring Assistant. Features dynamic keyword gap analysis, hallucination protection via original-document verification, and automatic LLM failover.

Claude Code skill that fact-checks social media post drafts and catches AI hallucinations (fabricated stats, laws, sources, quotes) before you publish.

Experimenting with AI Hallucinations, LLM learning shots

Executive-friendly RAG demo showing LLM hallucination vs grounded answers, plus bad RAG data vs good RAG data.

FlowMap is a Swift code graph and impact analysis tool for understanding and verifying AI-generated code changes.

Exploring which cat breeds are most frequently misclassified as dogs by CNN models.