4 ms·
The response would be more helpful if it directly addresses the arguments in posts from that search result.
by rs186 10mo ago
The response would be more helpful if it directly addresses the arguments in posts from that search result.
- maxspero 10mo agoThere are dozens of first generation AI detectors and they all suck. I'm not going to defend them. Most of them use perplexity based methods, which is a decent separators of AI and human text (80-90%) but has flaws that can't be overcome and high FPRs on ESL text. https://www.pangram.com/blog/why-perplexity-and-burstiness-fail-to-detect-ai https://www.pangram.com/blog/why-perplexity-and-burstiness-f... Pangram is fundamentally different technology, it's a large deep learning based model that is trained on hundreds of millions of human and AI examples. Some people see a dozen failed attempts at a problem as proof that the problem is impossible, but I would like to remind you that basically every major and minor technology was preceded by failed attempts.
- pixl97 10mo agoGAN.. Just feed the output of your algorithms back into the LLM while learning. At the end of the day the problem is impossible, but we're not there yet.
- QuadmasterXLII 10mo agoSome people see a dozen extremely profitable, extremely destructive attempts at a problem as proof that the problem is not a place for charitable interpretation.
- wordpad 10mo agoAnd you don't think a dozen of basically scams around the technology justify extreme scepticism?
- QuadmasterXLII 10mo agohuh?
- anonymouskimmer 10mo agoCan your software detect which LLMs most likely generated a text?
- maxspero 10mo agoPangram is trained on this task as well to add additional signal during training, but it's only ~90% accurate so we don't show the prediction in public-facing results