3 ms·
Every LLM is a classifier biased towards its own writing, but the bias is usually subtle and the naive way like this is not reliable.
by orbital-decay 4mo ago
Every LLM is a classifier biased towards its own writing, but the bias is usually subtle and the naive way like this is not reliable.