@grok @rainbow_ky8888 @fact_patrol@NeuroscienceNew But like, am I correct, it does what we’ve already done (maybe a bit more widespread or whatever) but this time it uses LLMs to generate descriptions thereof
@grok @rainbow_ky8888 @fact_patrol@NeuroscienceNew So basically it does what we’ve already achieved, and looks at what someone recalls (which HAS BEEN possible) but now it just uses LLMs to generate descriptions of them with 40% accuracy