Global News Post Fastest Global News Portal
Related Articles
- OpenAI has rolled again the ChatGPT 4o replace that led to the chatbot’s AI persona to be ‘sycophantic’
- CEO Sam Altman claims extra fixes are coming
- We may see a character variety possibility in ChatGPT someday
Following Sam Altman claiming ChatGPT 4o’s persona had change into ‘stressful’, OpenAI has now made adjustments to the AI style, rolling again closing week’s replace.
The corporate posted a weblog submit on its site highlighting the problem that had led to ChatGPT to change into extra ‘flattering or agreeable – continuously described as sycophantic.’
Users discovered the replace infuriating, with CEO Sam Altman sharing on X previous this week that the corporate was once operating on a repair for the ‘stressful’ behaviour.
the closing couple of GPT-4o updates have made the persona too sycophant-y and stressful (even supposing there are some excellent portions of it), and we’re operating on fixes asap, some nowadays and a few this week.in the future will percentage our learnings from this, it is been fascinating.April 27, 2025
In the weblog submit, OpenAI says, “In last week’s GPT‑4o update, we made adjustments aimed at improving the model’s default personality to make it feel more intuitive and effective across a variety of tasks.”
“However, in this update, we focused too much on short-term feedback, and did not fully account for how users’ interactions with ChatGPT evolve over time. As a result, GPT‑4o skewed towards responses that were overly supportive but disingenuous.”
In an effort to stop this from going down transferring ahead, the corporate has highlighted the way it seems to handle sycophancy someday.
For starters, coaching shall be used to ‘steer the style clear of sycophancy’ in addition to added guardrails installed position to ‘build up honesty and transparency’.
OpenAI additionally plans so as to add extra tactics for customers to supply comments on updates ahead of they’re rolled out to the broader public.
More persona updates to come back
we began rolling again the newest replace to GPT-4o closing evening it is now 100% rolled again at no cost customers and we will replace once more when it is completed for paid customers, with a bit of luck later nowadays we are operating on further fixes to style persona and can percentage extra within the coming daysApril 29, 2025
Following the rollback of the ChatGPT 4o replace, Altman wrote on X, “we’re working on additional fixes to model personality and will share more in the coming days.”
Putting two and two in combination, it seems like customers may get a ChatGPT persona variety instrument someday. OpenAI’s weblog submit mentions the corporate is “building new, easier ways for users to shape ChatGPT’s behavior” past customized directions.
If you wish to have to move additional and alter ChatGPT’s persona your self, we advise attempting the next recommended Reddit consumer TedHoliday:
“Can you please store a memory to permanently stop commenting on the quality and validity of my questions and simply get to the point? I don’t want to be manipulated with flattery by a robot, and I don’t want to have to skip past the garbage to get to the answer l’m looking for.”
You may additionally like
Source hyperlink
