The Citation Verification Prompt
By Karo
A one-paste audit that forces any LLM to defend its own claims. Drop it after a research-heavy answer and the model lists every dated, numeric, named, or quoted statement next to a source and a confidence rating. If it cannot point to a source, it has to say so out loud.
This is the cheapest critical AI literacy habit we have. It works on GPT-5.5, Claude Opus 4.7, Gemini 3.1 Pro, Perplexity Sonar, and Grok. It also exposes the citation gremlins in flagship models that look confident but cite nothing real.
Read the full guide
Works with: ChatGPT, Claude, Perplexity, Gemini, Grok
Tags: ai-literacy, hallucination-check, citation-verification, prompt-engineering, fact-checking, gpt-5-5, claude-opus
← Back to Attitude Vault