Ever struggle with a ML paper back in grad school and come away thinking this work is suss? Pass it through an LLM and $10 it finds 5 fatal errors.
Dan Roy claims LLMs find fatal research flaws for $10
Leo Boytsov warns that LLMs propagate literature biases
134054K
Original post unavailable.
Sentiment
Sentiment building, check back later.
Cluster Engagement
Digg Deeper
No Digg Deeper questions have been answered for this story yet.
Posts from X
Most Activity
Most Activity
VIEWS3.9KBOOKMARKS5LIKES35REPLIES1
True, but it also encodes some common biases in CS/ML, which it will NOT overcome on its own. With my favorite example of hash-function resampling (as it is done in Bloom-filter and LSH theory), it would work hard to convince you it's realistic. Even after pointing out the and explaining flaw, it says something along the lines: "you are right, the distribution is not the same", but the whole thing is nevertheless legit. 😅
Ever struggle with a ML paper back in grad school and come away thinking this work is suss? Pass it through an LLM and $10 it finds 5 fatal errors.