Worst Performers

Models Needing Improvement

These models showed the poorest performance in identifying fictional species, often claiming extensive knowledge about bacteria that don't exist. These models would benefit most from improved training on handling unknown entities.