AI Glossary
What is Hallucination?
Models generate plausible text, not verified facts — when the training data or context lacks an answer, they may produce a confident fabrication: fake citations, invented APIs, wrong numbers. Mitigations include RAG, asking models to cite sources, structured output validation, and human review for high-stakes use.
News momentum
Mentions in recent AI news titles and summaries, refreshed from the ingestion stream.
25
-3 vs prior 7d
- Hallucinations are a byproduct of the transformer model used in LLMs. Just really fancy plinko machines trying to land in the "please u" sq…
bluesky-ai · 19h
- Automatic Model Card Generation Using an LLM
arxiv-ai · 1d
- Lost in Speech: Trilingual Spoken Hallucination Detection Across Audio and Transcripts
arxiv-ai · 1d
- AI agents accessing email, calendars and personal data are useful, but one hallucination can delete an important email or send the wrong me…
bluesky-ai · 1d