This does not mean that if you have a loss function of predicting the next token from that data set, that the mind picked out by gradient…
Not checkable An opinion, value judgement or statement with nothing to check.
Not checkable: An opinion, value judgement or statement with nothing to check. Not scored.
Cite this claim
Eliezer Yudkowsky, “This does not mean that if you have a loss function of predicting the next token from that data set, that the mind picked out by gradient…” (said 30 March 2023). Verdict: not checkable. Reliability API, claim c_d230ea5300a3, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_d230ea5300a3
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
Verdicts are made automatically; how verdicts are made.