Metacognitive Boundaries in AI Model Evaluation

Comprising 60 examples designed to test model metacognition across 4 distinct reasoning tasks. It is used for evaluating artificial intelligence systems.

Use Cases

• Evaluate model performance on 4 distinct reasoning tasks to assess metacognitive boundaries.
• Benchmark AI systems using the 60 provided examples for metacognitive evaluation.
• Analyze model outputs across the specified tasks to identify patterns in metacognitive failure.

Strengths

• Contains 60 specific evaluation examples.
• Covers 4 distinct reasoning tasks for structured testing.

Limitations

• Small sample size of 60 examples limits statistical power.
• Unknown column structure and data format.
• Lacks metadata such as author, license, and update information.

Provenance

Source

Kaggle