9 ms·
This doesn‘t sound like AI. LLMs do not use dictionary definitions for translating a paper this way. It does look like some translation engine like Google Trans
by ivan_gammel 1mo ago
This doesn‘t sound like AI. LLMs do not use dictionary definitions for translating a paper this way. It does look like some translation engine like Google Translate or worse, maybe with some text obfuscator trying to avoid plagiarism detection.
- 1over137 1mo agoGoogle Translate uses an LLM, last I heard.
- kevin_thibedeau 1mo agoOnly since 2023.
- stymaar 1mo agoWDYM? It used an encoder-decoder transformer model long before 2023 (and also long before decoder-only transformers à la GPT became the new standard). Most people seem to believe LLM only describe these decoder-only architectures, but the term predate them by a big margin. BERT (encoder-only) was a large language model, and there were even language models before transformers were even a thing.
- scotty79 1mo agoIt switched very recently I think if at all. It was way worse at translations than Gemini even at the beginning of this year. Especially between a pair of languages other than English. It was mixing two completely different words that are the same in English.
- krige 1mo agoI've seen it make very LLM mistakes, esp. "hiccups" and palilalia this month.
- gmerc 1mo agoThat sounds like Anthrophics watermark explanation though.
- LPisGood 1mo agoAnthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it. The models always relied on some randomness, here the randomness is fixed and remembered for querying later.
- thaumasiotes 1mo ago> Anthropic’s watermark is just using a known collection of random seeds during output token generation. That’s it. Wouldn't that mean that checking the watermark... (1) Required knowing the prompt used to generate the text; and (2) was just as expensive as generating the text?
- hdjdbckcnfn 29d agowouldn't it mean reversing the output text into prompt and seed?
- rcxdude 1mo agoIt's gotta be a bit more than that, otherwise it would fail if you trimmed even one token off the start of the whole session, let alone removing the prompt entirely. (One approach that's been described is creating a bias in certain token pairs or short sequences: by hashing the previous n tokens, potentially with a secret seed, you create a list of essentially random allowed and disallowed next tokens, then bias the token selection to prefer allowed tokens, so long as they are plausible enough. Then you can check any long-enough subset of the output, if it deviates significantly enough from random chance on following this rule, it was almost certainly generated from a system following these rules.)
- wavemode 1mo agoLLMs are not the only form of AI.
- anamexis 1mo agoSimple substitution of phrases doesn’t sound like any AI.
- xgulfie 1mo agoAutocorrect unleashed doesn't sound like any AI either but that doesn't stop anyone from calling it that
- stymaar 1mo agoAren't most autocorrect transformer-based?
- sebastiennight 1mo agoHistorically, no. The simplest way to build autocorrect is to have a dictionary, look it up for each word, and suggest the "closest" dictionary word if a typed word is absent from the dictionary. It has existed way before transformers and on platforms that couldn't dream of running even the smallest transformer.
- sublinear 1mo agoWhat do you mean by "look it up" if not measure the closest dictionary word by edit distance? I think you'll find the broader idea of word vectorization to be remarkably similar to what text transformers do. https://en.wikipedia.org/wiki/Edit_distance https://en.wikipedia.org/wiki/Edit_distance
- sebastiennight 1mo agoYes, I meant, find the closest word by edit distance. I'm unsure where vectorization would factor in here?
- userbinator 1mo agoYes, this looks like something which predates AI by several decades: https://en.wikipedia.org/wiki/Article_spinning https://en.wikipedia.org/wiki/Article_spinning
- duskwuff 1mo agohttps://arxiv.org/abs/2107.06751 https://arxiv.org/abs/2107.06751 found that a lot of the outputs were characteristic of an online tool called SpinBot. (The tool no longer seems to work the same way that it did at the time.)
- Bombthecat 1mo agoMaybe deepl gone wild?
- pseudohadamard 1mo agoEither that or it was an AI trained on conversations between parents and teenagers.
- indrora 1mo agoWhen many of the core datasets were being produced, a lot of these companies mangled the text to be "friendlier" to try and avoid certain issues. The changes they made often resulted in this kind of weird nonsense: Swapping the word "Failure" for "Disappointment" in a vain attempt to keep it from saying "you're a failure", but also just stripping out words such as slurs and insults, resulting in at one point the hilarious situation in which Gemini for quite a while would INSIST that the lyrics for the Dire Straits song "Money for Nothing" contained "That little has his own jet airplane/that little is a millionaire" -- which in certain contexts changes the meaning of the song incredibly. These hack-and-slash text manglings have resulted in some of these clbuttic sort of writing styles being used during LLM-assisted writing. A more famous version of this was caused by one paper: https://world.edu/a-weird-phrase-is-plaguing-scientific-papers-and-we-traced-it-back-to-a-glitch-in-ai-training-data/ https://world.edu/a-weird-phrase-is-plaguing-scientific-pape...