The researchers used the Llama 2 13B large language model, incorporating their own parameters to construct "ChatBIT," a military-focused AI tool for intelligence processing.
"It's the first time there has been substantial evidence that PLA military experts have been systematically researching and trying to leverage the power of open-source LLMs, especially those of Meta, for military purposes," said Sunny Cheung.
ChatBIT was fine-tuned and optimized for dialogue and question-answering tasks in the military field, outperforming AI models nearly as capable as ChatGPT-4.
Meta claims the PLA is unauthorized to use the Llama model, while the Pentagon is closely monitoring competitors' advancements in AI capabilities.
Collection
[
|
...
]