Anthropic's Constitutional AI Approach

Anthropic developed 'Constitutional AI,' a method to train AI systems to be harmless and helpful by adhering to a set of written principles, reducing reliance on extensive human feedback.

Sources

• Toloka AI - Constitutional AI explained
• TDWI - Constitutional AI: How Anthropic Trains Models Using Written Principles
• On 'Constitutional' AI
• Anthropic - Constitutional AI: Harmlessness from AI Feedback
• Anthropic - Collective Constitutional AI: Aligning a Language Model with Public Input

Open the full topic