Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Translating Akkadian clay tablets with ChatGPT?

janromme.com

Post reply on HN
Print view View original

71–80 of 121 posts

Re: Translating Akkadian clay tablets with ChatGPT?

#71

Post by mdp2021 »

mdp2021 wrote: post #70
cuteboy19 wrote: post #40

Earlier quoted context omitted.

From 'do robots dream' to 'how do we make them stop'

> how do we make them stop The same way we do with humans: reflection, critical thinking. (Although, disrobing the "Cato the censor" robes, yes I loved this image.) (Although II, the original concept of "androids dreaming of electric sheep" is more related to the PKD theme of the human problem in terms of emotion, the deontic call for empathy, the human property of the inner trouble within a mechanistic perspective.…

...And I must add, back as the Censor:

the very fact that we quote Philip K. Dick with relevance to this context could suggest that:

hallucination is not the problem, the problem is further processing it to the desired state.

Re: Translating Akkadian clay tablets with ChatGPT?

#72

Post by namaria »

I worked for years as a translator (PT/EN). I hated 'just correcting' bad translations, from software or people bad at it. I cannot figure out why anyone would think generating half baked text or code with some generic tool and then have to parse it, identify errors and correct it would be preferable to just doing it right yourself. It's much less pleasant work.

Re: Translating Akkadian clay tablets with ChatGPT?

#73

Post by namaria »

jiggawatts wrote: post #50

Some people have been "strengthening" Chat GPT by re-running the same query several times, then combining the various outputs. I asked GPT 4 to translate one of the linked snippets twice , pasted in the original plus the variants, and told it to merge all three into an improved translation. I also gave it metadata and context from the page. Without that, it complained that an isolated snippet of Sumerian or Akkadian…

Averaging out bad output cannot produce good output. Or as the old saying goes, garbage in, garbage out.

Re: Translating Akkadian clay tablets with ChatGPT?

#74

Post by mschuster91 »

cuteboy19 wrote: post #40
ekidd wrote: post #30

Earlier quoted context omitted.

I can confirm that GPT 4 fails pretty badly on Middle Egyptian texts that are outside of its training set, using both MdC encoding and transliteration. I get translations which have at most a word or two in common with the original text. Almost pure hallucination.

From 'do robots dream' to 'how do we make them stop'

I think the breakthrough development from GPT to a general-purpose AI will be the introduction of at least some form of self-awareness.

Basically, at the moment, GPT models don't have the concept of "knowledge", or at least not in a way accessible from the inside of the model. Ask a human something they don't know and they'll tell you they don't know - GPT models will pull something out of thin air as they aren't aware they don't "know" the answer or that the input is insufficient.

Re: Translating Akkadian clay tablets with ChatGPT?

#75

Post by justsomehnguy »

bee_rider wrote: post #15

Could have gone without the Karen digression. It took up a third of the post, didn’t really have a point other than that he didn’t get a picture of the tablet, and wasn’t even really in the original spirit of the phrase anyway.

> It took up a third of the post

There would be no post if Karen didn't interfere fifteen years ago because author would had got his picture of the tablet.

Re: Translating Akkadian clay tablets with ChatGPT?

#76

Post by H8crilA »

Veen wrote: post #69
earthboundkid wrote: post #44

Anything is worth a shot, but 3.5 cannot consistently translate English into ROT13, so uh, corpus size matters.

That’s more a problem of tokenisation. In many cases, an LLM doesn’t see individual letters at all, so it can’t encode them one at a time.

I am quite sure GPT can undo tokenization if needed. It has no problem handling misspellings, plus you can just ask it to "spell things" L-I-K-E S-O.

Re: Translating Akkadian clay tablets with ChatGPT?

#77

Post by jiggawatts »

namaria wrote: post #73
jiggawatts wrote: post #50

Some people have been "strengthening" Chat GPT by re-running the same query several times, then combining the various outputs. I asked GPT 4 to translate one of the linked snippets twice , pasted in the original plus the variants, and told it to merge all three into an improved translation. I also gave it metadata and context from the page. Without that, it complained that an isolated snippet of Sumerian or Akkadian…

Averaging out bad output cannot produce good output. Or as the old saying goes, garbage in, garbage out.

It's not necessarily "garbage".

A known limitation of word-at-a-time GPT models is that they can't backtrack, edit, or iterate on what they're doing. You can see these limitations when asking them to solve puzzles involving "term substitution". They can only do these up to a certain complexity, and then they run out of steam.

However, if you run their output back through for a second round, they can continue substituting!

Essentially, GPT models make mistakes, but can also recognise those mistakes and fix them on subsequent iterations. Much like humans.

Another factor is that the "temperature" setting means that the outputs have an element to them that's purely random. Combining multiple outputs "anneals" out this randomness by converging on output that is the most consistent. Essentially, it's like asking a team of people to brainstorm, and then combining their partial solutions into a complete one. Once again, very human-like!

This has been tested (semi) scientifically. Running GPT 4 across some puzzles or exams might yield a 60% success rate if you give it one chance. If you let it solve each problem by running it four times and then letting it combine its own outputs itself (not via human-selection!), then it will get up to an 80% success rate. See: https://www.youtube.com/watch?v=wVzuvf9D9BU

Re: Translating Akkadian clay tablets with ChatGPT?

#78

Post by namaria »

jiggawatts wrote: post #77
namaria wrote: post #73

Earlier quoted context omitted.

Averaging out bad output cannot produce good output. Or as the old saying goes, garbage in, garbage out.

It's not necessarily "garbage". A known limitation of word-at-a-time GPT models is that they can't backtrack, edit, or iterate on what they're doing. You can see these limitations when asking them to solve puzzles involving "term substitution". They can only do these up to a certain complexity, and then they run out of steam. However, if you run their output back through for a second round, they can continue substitu…

There's nothing scientific about this. People are just holding a black box and saying "if I jiggle it like this, I can get better results". Approximating better results in simple contrived puzzles is hardly evidence that recycling the output makes it better in any absolute sense. Besides, translation being open ended and heavily context dependent will likely suffer from such recycling.

Re: Translating Akkadian clay tablets with ChatGPT?

#79

Post by jiggawatts »

namaria wrote: post #78
jiggawatts wrote: post #77

Earlier quoted context omitted.

It's not necessarily "garbage". A known limitation of word-at-a-time GPT models is that they can't backtrack, edit, or iterate on what they're doing. You can see these limitations when asking them to solve puzzles involving "term substitution". They can only do these up to a certain complexity, and then they run out of steam. However, if you run their output back through for a second round, they can continue substitu…

There's nothing scientific about this. People are just holding a black box and saying "if I jiggle it like this, I can get better results". Approximating better results in simple contrived puzzles is hardly evidence that recycling the output makes it better in any absolute sense. Besides, translation being open ended and heavily context dependent will likely suffer from such recycling.

They're taking objective measurements of experimental outcomes while varying aspects of the setup. How is that not scientific!?

"If I jiggle this is behaves differently" is like... half of all science!

At the end of the day, better is better, and I don't think that can be hand-waved away.

That's like saying someone is not smarter because they just "studied harder" and that doesn't count somehow.

Re: Translating Akkadian clay tablets with ChatGPT?

#80

Post by pishpash »

mschuster91 wrote: post #74
cuteboy19 wrote: post #40

Earlier quoted context omitted.

From 'do robots dream' to 'how do we make them stop'

I think the breakthrough development from GPT to a general-purpose AI will be the introduction of at least some form of self-awareness. Basically, at the moment, GPT models don't have the concept of "knowledge", or at least not in a way accessible from the inside of the model. Ask a human something they don't know and they'll tell you they don't know - GPT models will pull something out of thin air as they aren't awa…

ChatGPT already does this after alignment, it barely "knows" anything these days.
Post reply on HN