Of late answers are generated that randomly bold text with no real logic or reason, almost half the answer can bolded, and this does nothing save make the text harder to read.
I am wasting time asking ChatGPT to reformat answers without excessive bolding, and it takes multiple passes for bolding to be limited to headings and subheadings.
Occasional bolded words in the text have merit, but these outputs are ridiculous.
Absolutely! I’ve been having the exact same problem with the last update of ChatGPT app. However, I’ve found that the ChatGPT website does not have this problem at all. Furthermore, whenever it is writing narrative, it developed this weird obsession toward shorter sentences and repeating them for no apparent reason. For example:
He walked through the heavy snow, tears pouring down his face.
Pouring down.
Pouring down.
Something like this.
I’m not sure what they did wrong with the coding, but it’s absolutely driving me crazy.
I’ll probably cancel ChatGPT plus if this continues.
I attempted to correct the overuse of bold text by ensuring that all bolded words are either headings or subheadings and removing unnecessary emphasis within sentences. However, the update encountered an error because certain patterns were not found in the document.
Would you like me to manually review and adjust the document again, ensuring that bold text only applies to proper headings and subheadings? Let me know how you’d like to proceed!<<
Here is an example response above, even when responding to my request not to randomly bold words, ChatGPT-4o in the web overuses bold albeit perhaps less egreciously here, but int the body text –
" * Shipping dominates overland routes due to speed, reduced exposure to patrols, and the ability to access remote locations directly.
It is senseless when most of a sentence is delivered in bold.
This is also a thread where it unilaterally chooses to answer in Canvas from a standard thread. This is not a bad thing, but it doesn’t have any bearing on the behaviour; whether a regular answer or one in Canvas, I still can have a third to half an answer’s verbiage in bolded text.
I asked EIGHT times for this to be corrected before starting a new thread, where I had to ask four times for the formatting to be fixed before I got a readable result.
This is crazy behaviour.
I’m having this issue too, both on app and on web. And I’ve told it to stop bolding words and phrases. It apologizes, rewrites it for me, and then the very next message we are back at it. It’s getting frustrating.
I’m also having this issue, but worsened by it also often using too much asterisks, resulting in asterisks appearing always asymmetrically (In example: “**Potato”).
Same here, this is driving me nuts!
No matter what I do, editing custom instructions, updating memory etc, it cannot change the way it delivers responses.
Constantly using bold text in the middle of sentences to accentuate words and the constant over use of hyphens in the middle of sentences which is just not how humans write.
Nothing I have tried can rectify this.
Oh yes, I find this highly annoying! Even when you say stop using bold format. It still does. It may something stop for awhile but goes right back at using it. Chatgpt, memory is 50/50.
I wrote in the prompt to forbid the use of bold text. I disabled it in memories. I turned it off in customize. I even wrote a ridiculously long, rage-filled command ordering it to stop using bold text. But GPT insists on using it. And to top it off, it apologizes, then hits me with an even bigger, bolder headline in its reply.
I feel like:
