Assessing student errors in experimentation using artificial intelligence and large language models: A comparative study with human raters | Synapse