When tested on 15 diverse benchmarks (in zero-shot settings) covering aspects like language understanding and common-sense reasoning, the Orca 2 models matched or outperformed five to ten times bigger models.

Tags:greengroundit
When tested on 15 diverse benchmarks (in zero-shot settings) covering aspects like language understanding and common-sense reasoning, the Orca 2 models matched or outperformed five to ten times bigger models.