AI models can acquire backdoors from surprisingly few malicious documents
Anthropic study suggests "poison" training attacks don't scale with model size.
- Published date: 09-10-2025 10:03 PM
Anthropic study suggests "poison" training attacks don't scale with model size.