As a quick hack, is this flaw fixed by using top-p or top-k or min-p or whatever... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		arjvik 6 days ago \| parent \| context \| favorite \| on: Where Is GPT in the Chomsky Hierarchy? As a quick hack, is this flaw fixed by using top-p or top-k or min-p or whatever the current hottest logit sampling algorithm is?

FeepingCreature 6 days ago [–]

The "flaw" is also fixed by simply not recognizing the end-of-sampling token. Even the limited size of the context window can be worked around with tool calls.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact