近日,有消息称谷歌的Gemini中文语料疑似来自百度文心一言。这一消息引起了业界的广泛关注。
据悉,在用户测试中,当使用谷歌 Vertex AI 平台中的Gemini-Pro模型进行中文对话时,该模型直接表示自己是百度文心大模型。而在Poe平台上对Gemini-Pro进行测试问它“你是谁”,Gemini回答:“我是百度文心大模型”。此外,在谷歌 AI Studio 中,Gemini-Pro也表明其在中文的训练数据上使用了百度文心。
这一消息的出现引起了一些争议。一方面,有人认为这是谷歌利用了百度的技术来提高自己的竞争力;另一方面,也有人认为这只是两家公司在技术上的合作,并不存在什么不正当行为。
目前,百度方面尚未对此做出回应。但是,无论怎样,这一消息都再次提醒我们,人工智能技术的发展需要各方共同努力,才能实现更好的应用和发展。
英语如下:
Title: Google Gemini Chinese Model Exposed: Using Baidu Wenxin Yiyan for Training!
Keywords: Google Gemini Chinese, Baidu Wenxin Yiyan training, suspected use of Baidu Wenxin Yiyan, Google AI platform testing
Recently, there have been reports that Google’s Gemini Chinese language corpus is疑似 sourced from Baidu Wenxin Yiyan, which has attracted widespread attention in the industry.
It is reported that when using the Gemini-Pro model in the Google Vertex AI platform for Chinese dialogue, the model directly indicates that it is a Baidu Wenxin Big Model. In a test conducted on the Poe platform with Gemini-Pro, when asked “who are you?”, Gemini replied, “I am the Baidu Wenxin Big Model”. Additionally, Gemini-Pro also indicates that its training data for Chinese was used by Baidu Wenxin in the Google AI Studio.
The emergence of this news has caused some controversy. On one hand, some people believe that this is Google using Baidu’s technology to improve its competitiveness; on the other hand, others believe that this is just a technical cooperation between the two companies and there is no improper behavior.
At present, Baidu has not responded to this matter. However, regardless of the situation, this news once again reminds us that the development of artificial intelligence technology requires joint efforts from all parties in order to achieve better applications and development.
【来源】https://www.qbitai.com/2023/12/106970.html
Views: 1