GPT-4 is not getting worse • Hacker News Board

Re: GPT-4 is not getting worse

#51

Post by d0100 » Sat, Sep 16, 2023, 10:23 AM UTC

lijok wrote: post #37

rendall wrote: post #33

Earlier quoted context omitted.

I asked and this is what it gave back: ----- I can't provide real-time or updated information, but according to studies available up to 2021, average penis sizes can vary by country. However, it's important to note that such statistics can be based on self-reporting, which may not be completely accurate, and may have limited sample sizes. Here's a table based on available data: | Country | Average Length (Erect, cm)…

What they eating over there in brazil

Each other?

If not a case of self reported exaggeration, might be because we are incredibly miscegenated

Re: GPT-4 is not getting worse

#52

Post by practice9 » Sat, Sep 16, 2023, 10:26 AM UTC

Madmallard wrote: post #12

Theres at least a half dozen repeated queries with a 3+ month delay between in my chat history with chatGPT. The nerfing is that it does not try as hard to guess what you want and instead gives you a comment going “//fill this in” instead. This is with identical queries from February versus June or July. So that forces you at a minimum to query more. The other thing I noticed it doing is going “This is beyond the sco…

They biased the model to return shorter answers. Remember that "Continue" button if the reply was too long and cut off? I haven't seen it since they updated the model.

And to get a non-abridged version of the answer I need to ask it to provide "a detailed and comprehensive answer" more often.

Re: GPT-4 is not getting worse

#53

Post by dgellow » Sat, Sep 16, 2023, 10:27 AM UTC

Madmallard wrote: post #12

Theres at least a half dozen repeated queries with a 3+ month delay between in my chat history with chatGPT. The nerfing is that it does not try as hard to guess what you want and instead gives you a comment going “//fill this in” instead. This is with identical queries from February versus June or July. So that forces you at a minimum to query more. The other thing I noticed it doing is going “This is beyond the sco…

Exactly my experience. I cancelled my subscription a few weeks ago due to these issues, I now have to constantly follow answers with “please give me the complete answer, dont limit yourself to a shallow response”. Really frustrating to deal with. And I’m also spending way too much time pointing obvious mistakes…

Re: GPT-4 is not getting worse

#54

Post by dontupvoteme » Sat, Sep 16, 2023, 10:30 AM UTC

jw1224 wrote: post #35

There is an ongoing bug[1] with OpenAI's API, where it stops streaming responses after exactly 5 minutes. When I first came across the issue, I debugged it by writing a prompt along these lines: > Output every number from 1 to 10,000 in written form (e.g. "one", "two", etc.). Respond with one number per line in numeric order. As expected, the API would begin counting every number just as I asked. This would continue…

Wow, was this with the regular model (gpt-N) or a purported unchanging datestamped model like gpt-N-0314/0613?

Re: GPT-4 is not getting worse

#55

Post by muzani » Sat, Sep 16, 2023, 10:31 AM UTC

The gpt-4 API and ChatGPT-4 are basically different products. It's not clear which one this is comparing. Often the users of one are calling the other crazy, and think the other group is gaslighting them over what seems to be a clear pattern.

ChatGPT seems to fluctuate wildly in quality of expected output. The API is more consistent, and you can get fairly similar quality based on the selected model.

Re: GPT-4 is not getting worse

#56

Post by 15457345234 » Sat, Sep 16, 2023, 10:31 AM UTC

pyinstallwoes wrote: post #49

retube wrote: post #39

Earlier quoted context omitted.

How does chat-gpt actually get this right? This would appear to require some degree of reasoning, but as far as i understand its output is purely probabilistic, based on existing corpus of text.

Aren't your sentences probabilistically based on the existing corpus of text within your mind?

No? If that's how sentence forming worked it would be impossible to count backwards, or count at all for that matter.

Re: GPT-4 is not getting worse

#57

Post by 15457345234 » Sat, Sep 16, 2023, 10:34 AM UTC

1270018080 wrote: post #41

retube wrote: post #39

Earlier quoted context omitted.

How does chat-gpt actually get this right? This would appear to require some degree of reasoning, but as far as i understand its output is purely probabilistic, based on existing corpus of text.

You answered your own question.

Can you elaborate?

Re: GPT-4 is not getting worse

#58

Post by Stagnant » Sat, Sep 16, 2023, 10:34 AM UTC

jw1224 wrote: post #35

There is an ongoing bug[1] with OpenAI's API, where it stops streaming responses after exactly 5 minutes. When I first came across the issue, I debugged it by writing a prompt along these lines: > Output every number from 1 to 10,000 in written form (e.g. "one", "two", etc.). Respond with one number per line in numeric order. As expected, the API would begin counting every number just as I asked. This would continue…

Did you take in to account that the default GPT4 model changed in June? That would explain the response difference. You can still use the original model released in march with the GPT-4-0314 identifier.

Re: GPT-4 is not getting worse

#59

Post by wouldbecouldbe » Sat, Sep 16, 2023, 10:37 AM UTC

jw1224 wrote: post #35

There is an ongoing bug[1] with OpenAI's API, where it stops streaming responses after exactly 5 minutes. When I first came across the issue, I debugged it by writing a prompt along these lines: > Output every number from 1 to 10,000 in written form (e.g. "one", "two", etc.). Respond with one number per line in numeric order. As expected, the API would begin counting every number just as I asked. This would continue…

It's actually supper annoying.

I keep asking it to write out entire pieces of code, instead of only things it changed, can be annoying to merge it otherwise. But nowadays it acknowledges my request "I will write complete code" and proceed to still comment out parts.

Re: GPT-4 is not getting worse

#60

Post by stavros » Sat, Sep 16, 2023, 10:38 AM UTC

15457345234 wrote: post #56

pyinstallwoes wrote: post #49

Earlier quoted context omitted.

Aren't your sentences probabilistically based on the existing corpus of text within your mind?

No? If that's how sentence forming worked it would be impossible to count backwards, or count at all for that matter.

Me: Count backwards from ten

GPT-4: 10, 9, 8, 7, 6, 5, 4, 3, 2, 1

Me: Count backwards from a million to 999990

GPT-4: 1,000,000, 999,999, 999,998, 999,997, 999,996, 999,995, 999,994, 999,993, 999,992, 999,991, 999,990

I don't know about that.