TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

What's the Least Impressive Thing GPT-4 Won't Be Able to Do

18 点作者 dane-pgp超过 2 年前

7 条评论

nielpen超过 2 年前
Say &quot;I don&#x27;t know.&quot;<p>Back when GPT-3 was all the rage on the internet, I remember Kevin Lacker&#x27;s &quot;Giving GPT-3 a Turing Test&quot; left a real impression on me [1]. Short read, but gets to the crux of the issue. GPT-3 is a (really sophisticated) statistical model, GPT-4 will be no different. Maybe it will showcase new emergent properties of LLMs, like GPT-3&#x27;s in-prompt few-shot learning did. But the fundamental constraint of a statistical model optimized to a masking task remains -- they&#x27;re really bad at introspection and confidence-assessment.<p>[1] <a href="https:&#x2F;&#x2F;lacker.io&#x2F;ai&#x2F;2020&#x2F;07&#x2F;06&#x2F;giving-gpt-3-a-turing-test.html" rel="nofollow">https:&#x2F;&#x2F;lacker.io&#x2F;ai&#x2F;2020&#x2F;07&#x2F;06&#x2F;giving-gpt-3-a-turing-test.h...</a>
评论 #33910151 未加载
评论 #33910082 未加载
评论 #33912109 未加载
评论 #33909612 未加载
评论 #33911295 未加载
评论 #33909573 未加载
constantcrying超过 2 年前
Try asking ChatGPT about math:<p>&gt;What is 1111151111 * 100? Explain<p>In the mathematical operation that you have provided, the number 1111151111 is being multiplied by 100. When you multiply a number by 100, you simply add two zeroes to the end of the number. So in this case, the result of 1111151111 * 100 would be 11111511100.<p>To understand why this is the case, you can think about it in terms of place value. In the number 1111151111, the &quot;1&quot; in the far right position is worth 1, the &quot;1&quot; in the next position to the left is worth 10, the next &quot;1&quot; is worth 100, and so on. When you multiply a number by 100, you are essentially adding two zeroes to the end of the number, which means that you are effectively multiplying the value of each digit in the number by 100. So the &quot;1&quot; in the far right position would become 100, the &quot;1&quot; in the next position to the left would become 1000, and so on. This is why the result of 1111151111 * 100 is 11111511100.<p><i>It explains what to do and yet it still gets it wrong.</i><p>&gt;What is 1111151111 * 184582? Explain.<p>The product of 1111151111 and 184582 is 20748179353842. This can be calculated by performing the multiplication in the standard way, using the rules of arithmetic.<p>1111151111 x 184582 -------- 1111151111 2222302222 -------- 20748179353842<p>To calculate this product, we first multiply 1111151111 by 2 to get 2222302222. Then, we multiply 1111151111 by 82 to get 91678692462. Finally, we add these two products together to get 20748179353842. This is the final result of the multiplication.<p><i>This is pure fiction</i>
评论 #33910457 未加载
评论 #33915423 未加载
f154hfds超过 2 年前
A few prompts I just tried on chatGPT that it got wrong (not sure which one is the easiest):<p>Me: Hello chatGPT, I&#x27;m going to give you some prompts and you answer with a single word that seems to be the best fit, ok?<p>...<p>Me: Five is to three as blank is to seven<p>chatGPT: Two.<p>Me: The first prime after 90 is<p>chatGPT: 1.<p>Me: A cat&#x27;s lives times the number of days of the week is<p>chatGPT: seven.<p>------------------<p>A few that it got right that I thought were impressive:<p>Me: Grandma and grandpa had two sons, Harry and Harold. Harry had a daughter Tori. Harold is my dad, so Tori is my<p>chatGPT: cousin.<p>Me: Yesterday I walked on four, today I walk on two, tomorrow I walk on<p>chatGPT: three.<p>Me: house is to window as dogma is to<p>chatGPT: idea.<p>Me: Given the natural numbers, I remove all the even numbers and all the multiples of three. My first number remaining is<p>chatGPT: 1.
评论 #33915444 未加载
评论 #33915468 未加载
magneticnorth超过 2 年前
I&#x27;m enjoying trying some of these out on ChatGPT.<p>The prediction for multiplying two 5-digit numbers is correct for ChatGPT - the middle digits are wrong (non-deterministically).<p>The example &quot;Question: How many times does the word &quot;all&quot; occur in the following sentence: All the lice and all the mice were all very nice to all the mice and all the lice.&quot; results in answers of 5 or 6, about 50&#x2F;50 split.
评论 #33912765 未加载
armchairhacker超过 2 年前
Learning: These models, once trained, are essentially stateless: humans learn. If GPT-3 gets a question wrong you cannot permanently correct it, even a toddler adapts to its environment<p>A model like ChatGPT but it learns over time from user inputs (without just re-feeding prior inputs), and also just fine-tunes itself over time and learns more data, would be really impressive
dane-pgp超过 2 年前
I think this was actually a really thought-provoking question, when it was asked a few months ago, but now I&#x27;m wondering whether the guesses given have already been proven wrong by ChatGPT, or at least whether people would give guesses today that are very different from those given pre-ChatGPT.
benj111超过 2 年前
Boil an egg?