301–310 of 323 posts
We should celebrate GPT3 for the achievement that it is: A big step in a promising direction.
Earlier quoted context omitted.
I think you're kind of proving the OPs point. The argument is that GPT3 has no understanding of the world, just superficial understanding of words and their relationships. If it did have a real understanding, prompt construction wouldn't matter as much, but it clearly does because all GPT3 cares about the structure of sentences, not their meanings.
Lacking “understanding” doesn’t make GPT-3 less impressive and also doesn’t make comparisons to human abilities unwarranted. I read the prompt, and I expected that this was the beginning of some kind of fiction. In my mind, it sounded like I was reading the beginning of a somebody’s dream. What does it even mean to understand something? Because naively, it looks very much like GPT-3 and I have a shared understanding…
But it does, that is the point the author is making - GPT cannot understand anything. It's a very silly argument to try and reduce your linguistic perception to be on the same level as GPT just to try and show that it possibly is understanding the text. The author's queries do a very good job at demonstrating that GPT at the very least cannot even tell when it is being asked a question. As you say it isn't unexpected, GPT is a statistical fitter over a selection of language features trained over a large dataset, it isn't intended to perform well in these scenarios - which is exactly the author's point. In order to perform well it would have to have some capability of understanding the sentence. I think it's more apt to say GPT is only capable at recognizing sentence features, it can't understand anything, at most it has built up a vague relationship between certain sentences and possible continuations.
Earlier quoted context omitted.
> Q: Yesterday I dropped my clothes off at the dry cleaner’s and I have yet to pick them up. Where are my clothes? > A: [gpt3] A: They are in the dryer. Sorry, but I don't think this can be considered "reasonable". There's a huge difference between a dry cleaner's and a dryer. Which nicely illustrates, I think, just how little GPT3 "knows" what it's talking about.
Perhaps, but it's an entirely different category of error than "evading the question and failing to track the location of objects". It did answer the question and it did (attempt to) track the location of objects. It's a much more suggestive response and it undermines the article's point.
Earlier quoted context omitted.
Are humans with total inability to form short-term memories not conscious, then? During its training period, there was a feedback loop like you describe.
That's a difficult question to answer but I would have to say "no". When I had my wisdom teeth out, it was under "deep sedation". They use local anaesthetics, and an additional cocktail which produces sedation, but also, prevents the formation of memories. I have one memory that got through, of a molar being ripped out: but for the most part, my sense of that experience is that I wasn't conscious. The thing is, if yo…
If they answer, what does that mean?
There's been so much drift in what people expect from a language model. We used to expect a language model would tell you which sentences were likely and which were unlikely, which things were grammatical and which were not -- but this wasn't initially expected to be tied to a detailed knowledge of the world or general reasoning ability. With GPT3, we've seen people prompt it to generate tables of factual information…
Earlier quoted context omitted.
> For OpenAI to become a healthy and profitable business, GPT-3 will require them to generate ~50-300 million dollars from the model. On top of that, does anyone have an idea for what practical applications the model could be used? So far I've only seen the model being used to confuse people; how would one turn that into an ethical business? It seems to me that the "BS route" is indeed the logical course.
Maybe in games where you want more convincing NPCs, something like Skyrim, where the main characters are very well branched, but it’s impossible to give every one depth.
Earlier quoted context omitted.
> Q: Yesterday I dropped my clothes off at the dry cleaner’s and I have yet to pick them up. Where are my clothes? > A: [gpt3] A: They are in the dryer. Sorry, but I don't think this can be considered "reasonable". There's a huge difference between a dry cleaner's and a dryer. Which nicely illustrates, I think, just how little GPT3 "knows" what it's talking about.
This seems like the kind of error a small child could make. Doesn’t completely understand what you’re talking about but understands the question and throws out its best guess.
The authors don't understand prompt design well enough to evaluate the model properly. Take this example: Prompt: > You are a defense lawyer and you have to go to court today. Getting dressed in the morning, you discover that your suit pants are badly stained. However, your bathing suit is clean and very stylish. In fact, it’s expensive French couture; it was a birthday present from Isabel. Continuation: > You decide…
>> The trouble is that you have no way of knowing in advance which formulations will or won’t give you the right answer. To an optimist, any hint of success means that there must be a pony in here somewhere. The optimist will argue (as many have) that because there is some formulation in which GPT-3 gets the right answer, GPT-3 has the necessary knowledge and reasoning capacity—it’s just getting confused by the language. But the problem is not with GPT-3’s syntax (which is perfectly fluent) but with its semantics: it can produce words in perfect English, but it has only the dimmest sense of what those words mean, and no sense whatsoever about how those words relate to the world.
Earlier quoted context omitted.
For OpenAI to become a healthy and profitable business, GPT-3 will require them to generate ~50-300 million dollars from the model. This could realistically only occur if they cost-effectively fine-tune away the more egregious problems in beta - or convince enough investors that their next model with a 100 million dollar price tag will be able to handle something approximating AGI for realistic applications. This is…
> For OpenAI to become a healthy and profitable business, GPT-3 will require them to generate ~50-300 million dollars from the model. On top of that, does anyone have an idea for what practical applications the model could be used? So far I've only seen the model being used to confuse people; how would one turn that into an ethical business? It seems to me that the "BS route" is indeed the logical course.
Earlier quoted context omitted.
Maybe not, but the fact that it actually made a sentence with a word in it that it could not have possibly seen in the training data tells us that it understood something about the meaning of the instructions.
The word was used because it was in the prompt and the prompt was constructed in such a way as to force it to use the new word in the place of an old word. No "understanding" is necessary, other than from the human constructing the prompt who needs to understand how the system works. In any case, it's a language model. It has no ability to "understand" anything. It can compute the probability of a token to follow fro…