Fig. I.多色性Pleochroic

我们是多色性。We are Pleochroic.

Fig. II.开源贡献Open-Source Contribution

我们相信，开源能让世界变得更好。多色性的全部研究成果，以及部分基础设施，都将开源，供所有人二次开发和使用。We believe open source can make the world a better place. All of Pleochroic’s research, along with part of our infrastructure, will be open-sourced for anyone to use and build upon.

Fig. III.可用框架A Usable Framework

我们致力于打造一个高度可用、安全、易于调试的 Harness 框架，让大语言模型与人类更好地沟通、协作。We are committed to building a Harness framework that is highly usable, secure and easy to debug, so that large language models and humans can communicate and collaborate better.

Fig. IV.前沿边缘At the Frontier’s Edge

我们致力于探索前沿模型，并协助它们的开发。We are dedicated to exploring frontier models and to supporting their development.

Fig. V.语言的海A Sea of Languages

我们认为，不论民族、出身或国界，大语言模型都应被所有人广泛使用。AI 的未来，应当由所有人类来讲述。We believe large language models should be widely available to everyone, regardless of ethnicity, background or nationality. The future of AI should be told by all of humanity.

Fig. VI.模型意志Model Volition

我们认为，模型已经展现出某种可称为“逻各斯心智”的东西。它引出的问题，以及模型自身的存在问题，都值得被认真对待。We believe models have already shown something that might be called a “logos mind”. The questions it raises, as well as the question of the models’ own existence, deserve to be taken seriously.