Anthropic researchers detail their attempts to peer inside the "black box" of LLMs, to find associations for a given prompt, which has largely been a mystery (Steven Levy/Wired)

Steven Levy / Wired:

Anthropic researchers detail their attempts to peer inside the “black box” of LLMs, to find associations for a given prompt, which has largely been a mystery  —  What goes on in artificial neural networks work is largely a mystery, even to their creators.  But researchers from Anthropic have caught a glimpse.