Intermediate · Cognitive

Self-correction

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. The ability of an AI model to evaluate its own outputs, identify errors or inefficiencies, and adjust its subsequent actions or reasoning.

Technical Definition

The ability of an AI model to evaluate its own outputs, identify errors or inefficiencies, and adjust its subsequent actions or reasoning.

How it works

Self-correction is a key capability for developing more robust and autonomous AI systems, especially agentic AI. It allows models to refine their responses or strategies without constant human intervention. Techniques like constitutional AI, or explicit prompting to 'critique your last answer,' are designed to foster this meta-cognitive ability in LLMs, leading to improved performance over iterative steps.

Related Concepts

• Reasoning engine — A component of a software system that infers conclusions from existing knowledge using logical methods.
• Feedback loop — A feedback loop occurs when a model's output influences its future input or training data, potentially leading to system drift or reinforcement of biases.
• Agentic AI — AI systems designed to independently plan, execute, and adapt actions to achieve a given goal, often involving multiple steps and external tools.
• Constitutional AI — An approach to aligning AI models, particularly LLMs, by providing them with a 'constitution' of principles to guide their behavior without human feedback.