CINES virtual assistant

CINES provides a Large Language Model (LLM) based documentation helper. This assistant is imperfect and exhibits all the dubious and random behaviors one can expect in a LLM. We regularly finetune some of its features to try to reduce its imperfections. You can view this robot as a *useful idiot*. The robot’s view of the machine can be taken advantage of but should not be trusted. We trained the robot only on this documentation, no CINES user information is leaked during training. A fundamental problem with neural network based AI (as opposed to maybe, symbolic AI) is that if the output of such AI is obvious then you already knew the answer and, if it isn’t, the work you have to do to to prove it true or false is effectively unbounded. That said, it can still speed things up, as it is often easier to read and check than to produce.

Note

This robot is proudly brought to you using AMD hardware.

Warning

Once you have gained a bit of intuition through the robot, please, do fact check the answer using this documentation. Use this tool at your own risk.