7 ms·
Model fine tuning solves this. When (not if) the LA Times gets tired of paying writers, they will pay someone to tune an LLM using their corpus.
by grizzlychair 2y ago
Model fine tuning solves this. When (not if) the LA Times gets tired of paying writers, they will pay someone to tune an LLM using their corpus.
- santoshalper 2y agoThat only works because they have a corpus of high-quality human produced source material. What happens when they over farm the crops and ruin the soil? Perhaps they don't care, but should we?
- sterlind 2y agoI doubt it. LLMs seem to do a very superficial job generating summaries and reviews, likely because they don't know what to focus on (what information is relevant), and because they don't have the planning ability to structure a narrative. Instead, they tend to generate purple prose with buzzwords and cliches, filling in facts taken from the source material, and problems connecting clauses. It's a bit like image generation. It looks impressive at first glance, but generic, and when you focus you realize the image doesn't make sense, the details are wrong, the layout is confused. You can't generate working schematics or infographics with diffusion models - all you can make is filler content. Fine-tuning can give you different art styles, but it can't give you coherence. Fine-tuning can change an LLM's prose style, but it will still generate superficial, vacuous filler prose. Why would I pay the LA Times for drivel?