This particular implementation of the test seems flawed. The “invisible text” which would be read by a screen reader or visible if the problem statement were copied to some persistent scratch space while drafting a response would specify the inclusion of the word whether or not the author is an LLM.
Comments
This particular implementation of the test seems flawed. The “invisible text” which would be read by a screen reader or visible if the problem statement were copied to some persistent scratch space while drafting a response would specify the inclusion of the word whether or not the author is an LLM.
Yes because people are identical to llms and cannot see through that.