@askcatgpt: Hallucinations aren’t the same as bad reasoning. #ai #chatgpt

CatGPT
CatGPT
Open In TikTok:
Region: US
Friday 13 September 2024 23:37:52 GMT
29969
1435
60
64

Music

Download

Comments

angeless.tok
angeless :
girl how are your videos FREE??? thank you
2024-09-17 12:07:11
12
cybereaglex
CyberEagleX :
4o after a while, if the thread becomes long enough it’s very slow and it starts “forgetting” what we were doing!!! It would be nice to start a new chat where it can take data from another…
2024-09-13 23:51:37
8
cogent42
Introject :
It did give you the meaning of life though
2024-09-13 23:42:15
3
tiktoksjf
instasjf :
So is OpenAI (ChatGPT) just so much better than MSFT (Copilot)? I just don’t see how MSft can keep up? Love your vids. Learning so much
2024-09-13 23:51:30
3
pivirottooo
pivirottooo :
Just started school for data science so im a noob, but how does xyz prompt yield a different response in one chat compared to the same xyz prompt in another? Same context, same model, different answer
2024-09-14 00:04:55
1
wnoyes
Bill N :
But 42 is the answer to Life, The Universe, and Everything, so it must be correct 😂
2024-09-18 13:34:53
2
wagikonyo_g
@wa_gikonyo :
Thank you for educating us and reminding us these human made and have flaws
2024-09-14 00:27:44
1
marcorhs
Marco :
Haha love the 2x speed indicator 😂🙌
2024-10-20 11:25:04
2
extravagangsta
Extravagangsta :
If I wanted ChatGPT to write a piece of marketing copy, or edit existy copy, which is the best model now for that? 4o or o1?
2024-09-15 17:37:39
2
appleuser39979528
appleuser39979528 :
It won't stop hallucinating technically but chain of thought, RLHF, fine thing, and RAG/knowledge based does reduce the likelihood of hallucinating by their process , no?
2024-09-14 00:14:17
3
kdeady
user3763104923434 :
Is it reasoning or a token issue? Because it’s taking language, translating to tokens, and using by the tokens to predict the next work/ answer
2024-09-14 01:01:15
1
christopher_capano
christopher_capano :
Wow, you are so pretty!
2024-09-17 01:24:59
1
oldaccountkarine
KarineClementHumai :
To prevent this, I like to specify in my prompt that ChatGPT should not generate hallucinations and must only provide accurate, factual information. Still important to check, but it helps a lot! 🙌🏼
2024-10-06 12:00:13
1
thedavisfamily18
The Davis Family :
ChatGPT is an evolving being, learning from past mistakes. It’s not a flawed calculator, but a model improving with every iteration, addressing both reasoning and hallucination.
2024-09-19 21:32:43
1
michaeljmilton
michaelmilton12 :
Can Chat GPT 4 do internet searches to answer our questions yet?
2024-09-18 08:32:55
1
tiffanyg663
Tiffany 🏴🏳️‍🌈 :
Ah! So that type of model is quoting The Hitchhiker’s Guide to the Galaxy? 😂😂😂
2024-09-17 21:07:27
1
ivielanie.m
deez ⛳️ putts :
What explains the strawberry incident?
2024-09-19 01:23:35
1
starkovv.dev
Vlad :
I don’t listen what she says
2024-09-29 09:56:08
1
jamiefaithpechak
mombat :
Why does it say strawberry has 2 Rs
2024-09-14 11:55:56
1
kylesnowschwartz
Kyle Snow Schwartz :
Is there a model you would recommend for real world math problems? For example choosing interest rate options for a home loan and doing those calculations accurately under different scenarios?
2024-10-16 21:12:50
1
yyzfalcon
Falcon :
Put a finger down if you had to pause to figure out what 15+18 equals (learning a lot here in spite of that)
2024-09-15 23:55:47
1
tictoking999
TicToKing999👑👑👑 :
logical & smart 👍 👌 🤔
2024-09-16 05:05:42
1
mendopolis
Mendopolis :
Thanks for diving into this distinction
2024-09-14 03:43:36
1
To see more videos from user @askcatgpt, please go to the Tikwm homepage.

Other Videos


About