Comment on OpenAI Suspends ByteDance's Account After It Used GPT To Train Its Own AI Model

<- View Parent
ZickZack@fedia.io ⁨11⁩ ⁨months⁩ ago

Not necessarily: there have been recent works that indicate that filtering effects of fine tuned LLMs greatly improves the data efficiency (e.g phi-1).
Further, if you have e.g. human selection on top of LLM generated content you can get great results as the LLM generation can be used as a soft curriculum, with the human selection biasing towards higher quality.

source
Sort:hotnewtop