2 ms·
By now I suppose they could use an LLM to change the "personality" of the training data, then train a new LLM with it ;)
by amelius 2y ago
By now I suppose they could use an LLM to change the "personality" of the training data, then train a new LLM with it ;)
- nonethewiser 2y agoUgh. A derivative. We're in some ways already there. Not in terms of personality. But we're in a post-llm world. Training data contains some level of LLM generated material. I guess its on the model creators to ensure their data is good. But it seems like we might have a situation where the training material degrades over time. I imagine it being like if you apply a lossy compression algorithm to the same item many times. IE resaving a JPEG as JPEG. You lose data every time and it eventually becomes shit.